API tools faq
paste
Advertisement
Guest User

Untitled

a guest
Feb 21st, 2022
285
0
Never
Add comment
Not a member of Pastebin yet? Sign Up, it unlocks many cool features!
text 315.95 KB | None | 0 0
raw download clone embed print report
  1. C:\ngp\instant-ngp>cmake . -B build
  2. -- Building for: Visual Studio 16 2019
  3. -- Selecting Windows SDK version 10.0.19041.0 to target Windows 10.0.19044.
  4. -- The C compiler identification is MSVC 19.29.30140.0
  5. -- The CXX compiler identification is MSVC 19.29.30140.0
  6. -- The CUDA compiler identification is NVIDIA 11.6.55
  7. -- Detecting C compiler ABI info
  8. -- Detecting C compiler ABI info - done
  9. -- Check for working C compiler: C:/Program Files (x86)/Microsoft Visual Studio/2019/Community/VC/Tools/MSVC/14.29.30133/bin/Hostx64/x64/cl.exe - skipped
  10. -- Detecting C compile features
  11. -- Detecting C compile features - done
  12. -- Detecting CXX compiler ABI info
  13. -- Detecting CXX compiler ABI info - done
  14. -- Check for working CXX compiler: C:/Program Files (x86)/Microsoft Visual Studio/2019/Community/VC/Tools/MSVC/14.29.30133/bin/Hostx64/x64/cl.exe - skipped
  15. -- Detecting CXX compile features
  16. -- Detecting CXX compile features - done
  17. -- Detecting CUDA compiler ABI info
  18. -- Detecting CUDA compiler ABI info - done
  19. -- Check for working CUDA compiler: C:/Program Files/NVIDIA GPU Computing Toolkit/CUDA/v11.6/bin/nvcc.exe - skipped
  20. -- Detecting CUDA compile features
  21. -- Detecting CUDA compile features - done
  22. -- Looking for pthread.h
  23. -- Looking for pthread.h - not found
  24. -- Found Threads: TRUE
  25. -- Using Win32 for window creation
  26. -- Found OpenMP_C: -openmp (found version "2.0")
  27. -- Found OpenMP_CXX: -openmp (found version "2.0")
  28. -- Found OpenMP: TRUE (found version "2.0")
  29. -- OptiX_INSTALL_DIR value: C:\ProgramData\NVIDIA Corporation\OptiX SDK 7.4.0
  30. -- Found Python: C:/Users/alan/AppData/Local/Programs/Python/Python39/python.exe (found suitable version "3.9.10", minimum required is "3.7") found components: Interpreter Development Development.Module Development.Embed
  31. -- pybind11 v2.7.1
  32. CMake Warning (dev) at C:/Program Files/CMake/share/cmake-3.23/Modules/CMakeDependentOption.cmake:84 (message):
  33. Policy CMP0127 is not set: cmake_dependent_option() supports full Condition
  34. Syntax. Run "cmake --help-policy CMP0127" for policy details. Use the
  35. cmake_policy command to set the policy and suppress this warning.
  36. Call Stack (most recent call first):
  37. dependencies/pybind11/CMakeLists.txt:98 (cmake_dependent_option)
  38. This warning is for project developers. Use -Wno-dev to suppress it.
  39.  
  40. -- Performing Test HAS_MSVC_GL_LTCG
  41. -- Performing Test HAS_MSVC_GL_LTCG - Success
  42. -- Targeting GPU architectures: 75
  43. -- Configuring done
  44. -- Generating done
  45. -- Build files have been written to: C:/ngp/instant-ngp/build
  46.  
  47. C:\ngp\instant-ngp>cmake --build build --config RelWithDebInfo -j 16
  48. Microsoft (R) Build Engine version 16.11.2+f32259642 for .NET Framework
  49. Copyright (C) Microsoft Corporation. All rights reserved.
  50.  
  51. Checking Build System
  52. Building Custom Rule C:/ngp/instant-ngp/CMakeLists.txt
  53. Building Custom Rule C:/ngp/instant-ngp/dependencies/glfw/src/CMakeLists.txt
  54. context.c
  55. init.c
  56. input.c
  57. monitor.c
  58. vulkan.c
  59. window.c
  60. win32_init.c
  61. win32_joystick.c
  62. win32_monitor.c
  63. win32_time.c
  64. win32_thread.c
  65. win32_window.c
  66. wgl_context.c
  67. egl_context.c
  68. osmesa_context.c
  69. Generating Code...
  70. glfw_objects.vcxproj -> C:\ngp\instant-ngp\build\dependencies\glfw\src\glfw_objects.dir\RelWithDebInfo\glfw_objects.l
  71. ib
  72. Compiling CUDA source file ..\src\optix\pathescape.cu...
  73. Compiling CUDA source file ..\src\optix\raytrace.cu...
  74. Compiling CUDA source file ..\src\optix\raystab.cu...
  75.  
  76. C:\ngp\instant-ngp\build>"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\bin\nvcc.exe" -gencode=arch=comput
  77. e_52,code=\"sm_52,compute_52\" --use-local-env -ccbin "C:\Program Files (x86)\Microsoft Visual Studio\2019\Community\
  78. VC\Tools\MSVC\14.29.30133\bin\HostX64\x64" -x cu -I"C:\ngp\instant-ngp\dependencies" -I"C:\ngp\instant-ngp\dependen
  79. cies\eigen" -I"C:\ngp\instant-ngp\dependencies\filesystem" -I"C:\ngp\instant-ngp\dependencies\glfw\include" -I"C:\ngp
  80. \instant-ngp\dependencies\imgui\gl3w" -I"C:\ngp\instant-ngp\dependencies\nanovdb" -I"C:\ProgramData\NVIDIA Corporatio
  81. n\OptiX SDK 7.4.0\include" -I"C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include" -I"C:\ngp\instant-ngp\dependencie
  82. s\tiny-cuda-nn\dependencies" -I"C:\ngp\instant-ngp\dependencies\tinylogger" -I"C:\ngp\instant-ngp\include" -I"C:\ngp\
  83. instant-ngp\build" -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include" --keep-dir x64\RelWithDeb
  84. Info -maxrregcount=0 --machine 64 -ptx -cudart static --expt-relaxed-constexpr -std=c++14 -Xcompiler="/EHsc -Zi -Ob
  85. 1" -D_WINDOWS -DNDEBUG -DTCNN_MIN_GPU_ARCH=75 -DTCNN_SHAMPOO -DNGP_GUI -DNGP_OPTIX -D"NGP_VERSION=\"1.0dev\"" -D"CMAK
  86. E_INTDIR=\"RelWithDebInfo\"" -D_MBCS -D"CMAKE_INTDIR=\"RelWithDebInfo\"" -o optix_program.dir\RelWithDebInfo\pathesca
  87. pe.ptx "C:\ngp\instant-ngp\src\optix\pathescape.cu"
  88.  
  89. C:\ngp\instant-ngp\build>"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\bin\nvcc.exe" -gencode=arch=comput
  90. e_52,code=\"sm_52,compute_52\" --use-local-env -ccbin "C:\Program Files (x86)\Microsoft Visual Studio\2019\Community\
  91. VC\Tools\MSVC\14.29.30133\bin\HostX64\x64" -x cu -I"C:\ngp\instant-ngp\dependencies" -I"C:\ngp\instant-ngp\dependen
  92. cies\eigen" -I"C:\ngp\instant-ngp\dependencies\filesystem" -I"C:\ngp\instant-ngp\dependencies\glfw\include" -I"C:\ngp
  93. \instant-ngp\dependencies\imgui\gl3w" -I"C:\ngp\instant-ngp\dependencies\nanovdb" -I"C:\ProgramData\NVIDIA Corporatio
  94. n\OptiX SDK 7.4.0\include" -I"C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include" -I"C:\ngp\instant-ngp\dependencie
  95. s\tiny-cuda-nn\dependencies" -I"C:\ngp\instant-ngp\dependencies\tinylogger" -I"C:\ngp\instant-ngp\include" -I"C:\ngp\
  96. instant-ngp\build" -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include" --keep-dir x64\RelWithDeb
  97. Info -maxrregcount=0 --machine 64 -ptx -cudart static --expt-relaxed-constexpr -std=c++14 -Xcompiler="/EHsc -Zi -Ob
  98. 1" -D_WINDOWS -DNDEBUG -DTCNN_MIN_GPU_ARCH=75 -DTCNN_SHAMPOO -DNGP_GUI -DNGP_OPTIX -D"NGP_VERSION=\"1.0dev\"" -D"CMAK
  99. E_INTDIR=\"RelWithDebInfo\"" -D_MBCS -D"CMAKE_INTDIR=\"RelWithDebInfo\"" -o optix_program.dir\RelWithDebInfo\raystab.
  100. ptx "C:\ngp\instant-ngp\src\optix\raystab.cu"
  101.  
  102. C:\ngp\instant-ngp\build>"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\bin\nvcc.exe" -gencode=arch=comput
  103. e_52,code=\"sm_52,compute_52\" --use-local-env -ccbin "C:\Program Files (x86)\Microsoft Visual Studio\2019\Community\
  104. VC\Tools\MSVC\14.29.30133\bin\HostX64\x64" -x cu -I"C:\ngp\instant-ngp\dependencies" -I"C:\ngp\instant-ngp\dependen
  105. cies\eigen" -I"C:\ngp\instant-ngp\dependencies\filesystem" -I"C:\ngp\instant-ngp\dependencies\glfw\include" -I"C:\ngp
  106. \instant-ngp\dependencies\imgui\gl3w" -I"C:\ngp\instant-ngp\dependencies\nanovdb" -I"C:\ProgramData\NVIDIA Corporatio
  107. n\OptiX SDK 7.4.0\include" -I"C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include" -I"C:\ngp\instant-ngp\dependencie
  108. s\tiny-cuda-nn\dependencies" -I"C:\ngp\instant-ngp\dependencies\tinylogger" -I"C:\ngp\instant-ngp\include" -I"C:\ngp\
  109. instant-ngp\build" -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include" --keep-dir x64\RelWithDeb
  110. Info -maxrregcount=0 --machine 64 -ptx -cudart static --expt-relaxed-constexpr -std=c++14 -Xcompiler="/EHsc -Zi -Ob
  111. 1" -D_WINDOWS -DNDEBUG -DTCNN_MIN_GPU_ARCH=75 -DTCNN_SHAMPOO -DNGP_GUI -DNGP_OPTIX -D"NGP_VERSION=\"1.0dev\"" -D"CMAK
  112. E_INTDIR=\"RelWithDebInfo\"" -D_MBCS -D"CMAKE_INTDIR=\"RelWithDebInfo\"" -o optix_program.dir\RelWithDebInfo\raytrace
  113. .ptx "C:\ngp\instant-ngp\src\optix\raytrace.cu"
  114. raystab.cu
  115. raytrace.cu
  116. pathescape.cu
  117. Building Custom Rule C:/ngp/instant-ngp/dependencies/tiny-cuda-nn/src/CMakeLists.txt
  118. Compiling CUDA source file ..\..\..\..\dependencies\tiny-cuda-nn\src\cpp_api.cu...
  119. Compiling CUDA source file ..\..\..\..\dependencies\tiny-cuda-nn\src\cutlass_mlp.cu...
  120. Compiling CUDA source file ..\..\..\..\dependencies\tiny-cuda-nn\src\common_device.cu...
  121. Compiling CUDA source file ..\..\..\..\dependencies\tiny-cuda-nn\src\common.cu...
  122. Compiling CUDA source file ..\..\..\..\dependencies\tiny-cuda-nn\src\object.cu...
  123. Compiling CUDA source file ..\..\..\..\dependencies\tiny-cuda-nn\src\encoding.cu...
  124. Compiling CUDA source file ..\..\..\..\dependencies\tiny-cuda-nn\src\optimizer.cu...
  125. Compiling CUDA source file ..\..\..\..\dependencies\tiny-cuda-nn\src\cutlass_resnet.cu...
  126. Compiling CUDA source file ..\..\..\..\dependencies\tiny-cuda-nn\src\network.cu...
  127. Compiling CUDA source file ..\..\..\..\dependencies\tiny-cuda-nn\src\reduce_sum.cu...
  128. Compiling CUDA source file ..\..\..\..\dependencies\tiny-cuda-nn\src\loss.cu...
  129. Compiling CUDA source file ..\..\..\..\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu...
  130.  
  131. C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src>"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\bin\
  132. nvcc.exe" -gencode=arch=compute_52,code=\"sm_52,compute_52\" --use-local-env -ccbin "C:\Program Files (x86)\Microsoft
  133. Visual Studio\2019\Community\VC\Tools\MSVC\14.29.30133\bin\HostX64\x64" -x cu -I"C:\ngp\instant-ngp\dependencies"
  134. -I"C:\ngp\instant-ngp\dependencies\eigen" -I"C:\ngp\instant-ngp\dependencies\filesystem" -I"C:\ngp\instant-ngp\depend
  135. encies\glfw\include" -I"C:\ngp\instant-ngp\dependencies\imgui\gl3w" -I"C:\ngp\instant-ngp\dependencies\nanovdb" -I"C:
  136. \ProgramData\NVIDIA Corporation\OptiX SDK 7.4.0\include" -I"C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include" -I"
  137. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\dependencies" -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.
  138. 6\include" --keep-dir x64\RelWithDebInfo -maxrregcount=0 --machine 64 --compile -cudart static --extended-lambd
  139. a --expt-relaxed-constexpr -std=c++14 -Xcompiler="/EHsc -Zi -Ob1 -bigobj" -D_WINDOWS -DNDEBUG -DNGP_GUI -DNGP_OPTIX
  140. -DTCNN_MIN_GPU_ARCH=75 -DTCNN_SHAMPOO -D"CMAKE_INTDIR=\"RelWithDebInfo\"" -D_MBCS -D"CMAKE_INTDIR=\"RelWithDebInfo\"
  141. " -Xcompiler "/EHsc /W1 /nologo /O2 /FdC:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\RelWithDebInfo\tiny-cud
  142. a-nn.pdb /FS /Zi /MD /GR" -o tiny-cuda-nn.dir\RelWithDebInfo\cutlass_mlp.obj "C:\ngp\instant-ngp\dependencies\tiny-c
  143. uda-nn\src\cutlass_mlp.cu"
  144.  
  145. C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src>"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\bin\
  146. nvcc.exe" -gencode=arch=compute_52,code=\"sm_52,compute_52\" --use-local-env -ccbin "C:\Program Files (x86)\Microsoft
  147. Visual Studio\2019\Community\VC\Tools\MSVC\14.29.30133\bin\HostX64\x64" -x cu -I"C:\ngp\instant-ngp\dependencies"
  148. -I"C:\ngp\instant-ngp\dependencies\eigen" -I"C:\ngp\instant-ngp\dependencies\filesystem" -I"C:\ngp\instant-ngp\depend
  149. encies\glfw\include" -I"C:\ngp\instant-ngp\dependencies\imgui\gl3w" -I"C:\ngp\instant-ngp\dependencies\nanovdb" -I"C:
  150. \ProgramData\NVIDIA Corporation\OptiX SDK 7.4.0\include" -I"C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include" -I"
  151. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\dependencies" -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.
  152. 6\include" --keep-dir x64\RelWithDebInfo -maxrregcount=0 --machine 64 --compile -cudart static --extended-lambd
  153. a --expt-relaxed-constexpr -std=c++14 -Xcompiler="/EHsc -Zi -Ob1 -bigobj" -D_WINDOWS -DNDEBUG -DNGP_GUI -DNGP_OPTIX
  154. -DTCNN_MIN_GPU_ARCH=75 -DTCNN_SHAMPOO -D"CMAKE_INTDIR=\"RelWithDebInfo\"" -D_MBCS -D"CMAKE_INTDIR=\"RelWithDebInfo\"
  155. " -Xcompiler "/EHsc /W1 /nologo /O2 /FdC:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\RelWithDebInfo\tiny-cud
  156. a-nn.pdb /FS /Zi /MD /GR" -o tiny-cuda-nn.dir\RelWithDebInfo\cpp_api.obj "C:\ngp\instant-ngp\dependencies\tiny-cuda-
  157. nn\src\cpp_api.cu"
  158.  
  159. C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src>"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\bin\
  160. nvcc.exe" -gencode=arch=compute_52,code=\"sm_52,compute_52\" --use-local-env -ccbin "C:\Program Files (x86)\Microsoft
  161. Visual Studio\2019\Community\VC\Tools\MSVC\14.29.30133\bin\HostX64\x64" -x cu -I"C:\ngp\instant-ngp\dependencies"
  162. -I"C:\ngp\instant-ngp\dependencies\eigen" -I"C:\ngp\instant-ngp\dependencies\filesystem" -I"C:\ngp\instant-ngp\depend
  163. encies\glfw\include" -I"C:\ngp\instant-ngp\dependencies\imgui\gl3w" -I"C:\ngp\instant-ngp\dependencies\nanovdb" -I"C:
  164. \ProgramData\NVIDIA Corporation\OptiX SDK 7.4.0\include" -I"C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include" -I"
  165. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\dependencies" -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.
  166. 6\include" --keep-dir x64\RelWithDebInfo -maxrregcount=0 --machine 64 --compile -cudart static --extended-lambd
  167. a --expt-relaxed-constexpr -std=c++14 -Xcompiler="/EHsc -Zi -Ob1 -bigobj" -D_WINDOWS -DNDEBUG -DNGP_GUI -DNGP_OPTIX
  168. -DTCNN_MIN_GPU_ARCH=75 -DTCNN_SHAMPOO -D"CMAKE_INTDIR=\"RelWithDebInfo\"" -D_MBCS -D"CMAKE_INTDIR=\"RelWithDebInfo\"
  169. " -Xcompiler "/EHsc /W1 /nologo /O2 /FdC:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\RelWithDebInfo\tiny-cud
  170. a-nn.pdb /FS /Zi /MD /GR" -o tiny-cuda-nn.dir\RelWithDebInfo\common.obj "C:\ngp\instant-ngp\dependencies\tiny-cuda-n
  171. n\src\common.cu"
  172.  
  173. C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src>"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\bin\
  174. nvcc.exe" -gencode=arch=compute_52,code=\"sm_52,compute_52\" --use-local-env -ccbin "C:\Program Files (x86)\Microsoft
  175. Visual Studio\2019\Community\VC\Tools\MSVC\14.29.30133\bin\HostX64\x64" -x cu -I"C:\ngp\instant-ngp\dependencies"
  176. -I"C:\ngp\instant-ngp\dependencies\eigen" -I"C:\ngp\instant-ngp\dependencies\filesystem" -I"C:\ngp\instant-ngp\depend
  177. encies\glfw\include" -I"C:\ngp\instant-ngp\dependencies\imgui\gl3w" -I"C:\ngp\instant-ngp\dependencies\nanovdb" -I"C:
  178. \ProgramData\NVIDIA Corporation\OptiX SDK 7.4.0\include" -I"C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include" -I"
  179. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\dependencies" -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.
  180. 6\include" --keep-dir x64\RelWithDebInfo -maxrregcount=0 --machine 64 --compile -cudart static --extended-lambd
  181. a --expt-relaxed-constexpr -std=c++14 -Xcompiler="/EHsc -Zi -Ob1 -bigobj" -D_WINDOWS -DNDEBUG -DNGP_GUI -DNGP_OPTIX
  182. -DTCNN_MIN_GPU_ARCH=75 -DTCNN_SHAMPOO -D"CMAKE_INTDIR=\"RelWithDebInfo\"" -D_MBCS -D"CMAKE_INTDIR=\"RelWithDebInfo\"
  183. " -Xcompiler "/EHsc /W1 /nologo /O2 /FdC:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\RelWithDebInfo\tiny-cud
  184. a-nn.pdb /FS /Zi /MD /GR" -o tiny-cuda-nn.dir\RelWithDebInfo\common_device.obj "C:\ngp\instant-ngp\dependencies\tiny
  185. -cuda-nn\src\common_device.cu"
  186.  
  187. C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src>"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\bin\
  188. nvcc.exe" -gencode=arch=compute_52,code=\"sm_52,compute_52\" --use-local-env -ccbin "C:\Program Files (x86)\Microsoft
  189. Visual Studio\2019\Community\VC\Tools\MSVC\14.29.30133\bin\HostX64\x64" -x cu -I"C:\ngp\instant-ngp\dependencies"
  190. -I"C:\ngp\instant-ngp\dependencies\eigen" -I"C:\ngp\instant-ngp\dependencies\filesystem" -I"C:\ngp\instant-ngp\depend
  191. encies\glfw\include" -I"C:\ngp\instant-ngp\dependencies\imgui\gl3w" -I"C:\ngp\instant-ngp\dependencies\nanovdb" -I"C:
  192. \ProgramData\NVIDIA Corporation\OptiX SDK 7.4.0\include" -I"C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include" -I"
  193. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\dependencies" -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.
  194. 6\include" --keep-dir x64\RelWithDebInfo -maxrregcount=0 --machine 64 --compile -cudart static --extended-lambd
  195. a --expt-relaxed-constexpr -std=c++14 -Xcompiler="/EHsc -Zi -Ob1 -bigobj" -D_WINDOWS -DNDEBUG -DNGP_GUI -DNGP_OPTIX
  196. -DTCNN_MIN_GPU_ARCH=75 -DTCNN_SHAMPOO -D"CMAKE_INTDIR=\"RelWithDebInfo\"" -D_MBCS -D"CMAKE_INTDIR=\"RelWithDebInfo\"
  197. " -Xcompiler "/EHsc /W1 /nologo /O2 /FdC:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\RelWithDebInfo\tiny-cud
  198. a-nn.pdb /FS /Zi /MD /GR" -o tiny-cuda-nn.dir\RelWithDebInfo\reduce_sum.obj "C:\ngp\instant-ngp\dependencies\tiny-cu
  199. da-nn\src\reduce_sum.cu"
  200.  
  201. C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src>"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\bin\
  202. nvcc.exe" -gencode=arch=compute_52,code=\"sm_52,compute_52\" --use-local-env -ccbin "C:\Program Files (x86)\Microsoft
  203. Visual Studio\2019\Community\VC\Tools\MSVC\14.29.30133\bin\HostX64\x64" -x cu -I"C:\ngp\instant-ngp\dependencies"
  204. -I"C:\ngp\instant-ngp\dependencies\eigen" -I"C:\ngp\instant-ngp\dependencies\filesystem" -I"C:\ngp\instant-ngp\depend
  205. encies\glfw\include" -I"C:\ngp\instant-ngp\dependencies\imgui\gl3w" -I"C:\ngp\instant-ngp\dependencies\nanovdb" -I"C:
  206. \ProgramData\NVIDIA Corporation\OptiX SDK 7.4.0\include" -I"C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include" -I"
  207. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\dependencies" -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.
  208. 6\include" --keep-dir x64\RelWithDebInfo -maxrregcount=0 --machine 64 --compile -cudart static --extended-lambd
  209. a --expt-relaxed-constexpr -std=c++14 -Xcompiler="/EHsc -Zi -Ob1 -bigobj" -D_WINDOWS -DNDEBUG -DNGP_GUI -DNGP_OPTIX
  210. -DTCNN_MIN_GPU_ARCH=75 -DTCNN_SHAMPOO -D"CMAKE_INTDIR=\"RelWithDebInfo\"" -D_MBCS -D"CMAKE_INTDIR=\"RelWithDebInfo\"
  211. " -Xcompiler "/EHsc /W1 /nologo /O2 /FdC:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\RelWithDebInfo\tiny-cud
  212. a-nn.pdb /FS /Zi /MD /GR" -o tiny-cuda-nn.dir\RelWithDebInfo\loss.obj "C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\
  213. src\loss.cu"
  214.  
  215. C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src>"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\bin\
  216. nvcc.exe" -gencode=arch=compute_52,code=\"sm_52,compute_52\" --use-local-env -ccbin "C:\Program Files (x86)\Microsoft
  217. Visual Studio\2019\Community\VC\Tools\MSVC\14.29.30133\bin\HostX64\x64" -x cu -I"C:\ngp\instant-ngp\dependencies"
  218. -I"C:\ngp\instant-ngp\dependencies\eigen" -I"C:\ngp\instant-ngp\dependencies\filesystem" -I"C:\ngp\instant-ngp\depend
  219. encies\glfw\include" -I"C:\ngp\instant-ngp\dependencies\imgui\gl3w" -I"C:\ngp\instant-ngp\dependencies\nanovdb" -I"C:
  220. \ProgramData\NVIDIA Corporation\OptiX SDK 7.4.0\include" -I"C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include" -I"
  221. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\dependencies" -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.
  222. 6\include" --keep-dir x64\RelWithDebInfo -maxrregcount=0 --machine 64 --compile -cudart static --extended-lambd
  223. a --expt-relaxed-constexpr -std=c++14 -Xcompiler="/EHsc -Zi -Ob1 -bigobj" -D_WINDOWS -DNDEBUG -DNGP_GUI -DNGP_OPTIX
  224. -DTCNN_MIN_GPU_ARCH=75 -DTCNN_SHAMPOO -D"CMAKE_INTDIR=\"RelWithDebInfo\"" -D_MBCS -D"CMAKE_INTDIR=\"RelWithDebInfo\"
  225. " -Xcompiler "/EHsc /W1 /nologo /O2 /FdC:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\RelWithDebInfo\tiny-cud
  226. a-nn.pdb /FS /Zi /MD /GR" -o tiny-cuda-nn.dir\RelWithDebInfo\network.obj "C:\ngp\instant-ngp\dependencies\tiny-cuda-
  227. nn\src\network.cu"
  228.  
  229. C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src>"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\bin\
  230. nvcc.exe" -gencode=arch=compute_52,code=\"sm_52,compute_52\" --use-local-env -ccbin "C:\Program Files (x86)\Microsoft
  231. Visual Studio\2019\Community\VC\Tools\MSVC\14.29.30133\bin\HostX64\x64" -x cu -I"C:\ngp\instant-ngp\dependencies"
  232. -I"C:\ngp\instant-ngp\dependencies\eigen" -I"C:\ngp\instant-ngp\dependencies\filesystem" -I"C:\ngp\instant-ngp\depend
  233. encies\glfw\include" -I"C:\ngp\instant-ngp\dependencies\imgui\gl3w" -I"C:\ngp\instant-ngp\dependencies\nanovdb" -I"C:
  234. \ProgramData\NVIDIA Corporation\OptiX SDK 7.4.0\include" -I"C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include" -I"
  235. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\dependencies" -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.
  236. 6\include" --keep-dir x64\RelWithDebInfo -maxrregcount=0 --machine 64 --compile -cudart static --extended-lambd
  237. a --expt-relaxed-constexpr -std=c++14 -Xcompiler="/EHsc -Zi -Ob1 -bigobj" -D_WINDOWS -DNDEBUG -DNGP_GUI -DNGP_OPTIX
  238. -DTCNN_MIN_GPU_ARCH=75 -DTCNN_SHAMPOO -D"CMAKE_INTDIR=\"RelWithDebInfo\"" -D_MBCS -D"CMAKE_INTDIR=\"RelWithDebInfo\"
  239. " -Xcompiler "/EHsc /W1 /nologo /O2 /FdC:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\RelWithDebInfo\tiny-cud
  240. a-nn.pdb /FS /Zi /MD /GR" -o tiny-cuda-nn.dir\RelWithDebInfo\cutlass_resnet.obj "C:\ngp\instant-ngp\dependencies\tin
  241. y-cuda-nn\src\cutlass_resnet.cu"
  242.  
  243. C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src>"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\bin\
  244. nvcc.exe" -gencode=arch=compute_52,code=\"sm_52,compute_52\" --use-local-env -ccbin "C:\Program Files (x86)\Microsoft
  245. Visual Studio\2019\Community\VC\Tools\MSVC\14.29.30133\bin\HostX64\x64" -x cu -I"C:\ngp\instant-ngp\dependencies"
  246. -I"C:\ngp\instant-ngp\dependencies\eigen" -I"C:\ngp\instant-ngp\dependencies\filesystem" -I"C:\ngp\instant-ngp\depend
  247. encies\glfw\include" -I"C:\ngp\instant-ngp\dependencies\imgui\gl3w" -I"C:\ngp\instant-ngp\dependencies\nanovdb" -I"C:
  248. \ProgramData\NVIDIA Corporation\OptiX SDK 7.4.0\include" -I"C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include" -I"
  249. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\dependencies" -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.
  250. 6\include" --keep-dir x64\RelWithDebInfo -maxrregcount=0 --machine 64 --compile -cudart static --extended-lambd
  251. a --expt-relaxed-constexpr -std=c++14 -Xcompiler="/EHsc -Zi -Ob1 -bigobj" -D_WINDOWS -DNDEBUG -DNGP_GUI -DNGP_OPTIX
  252. -DTCNN_MIN_GPU_ARCH=75 -DTCNN_SHAMPOO -D"CMAKE_INTDIR=\"RelWithDebInfo\"" -D_MBCS -D"CMAKE_INTDIR=\"RelWithDebInfo\"
  253. " -Xcompiler "/EHsc /W1 /nologo /O2 /FdC:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\RelWithDebInfo\tiny-cud
  254. a-nn.pdb /FS /Zi /MD /GR" -o tiny-cuda-nn.dir\RelWithDebInfo\optimizer.obj "C:\ngp\instant-ngp\dependencies\tiny-cud
  255. a-nn\src\optimizer.cu"
  256.  
  257. C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src>"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\bin\
  258. nvcc.exe" -gencode=arch=compute_52,code=\"sm_52,compute_52\" --use-local-env -ccbin "C:\Program Files (x86)\Microsoft
  259. Visual Studio\2019\Community\VC\Tools\MSVC\14.29.30133\bin\HostX64\x64" -x cu -I"C:\ngp\instant-ngp\dependencies"
  260. -I"C:\ngp\instant-ngp\dependencies\eigen" -I"C:\ngp\instant-ngp\dependencies\filesystem" -I"C:\ngp\instant-ngp\depend
  261. encies\glfw\include" -I"C:\ngp\instant-ngp\dependencies\imgui\gl3w" -I"C:\ngp\instant-ngp\dependencies\nanovdb" -I"C:
  262. \ProgramData\NVIDIA Corporation\OptiX SDK 7.4.0\include" -I"C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include" -I"
  263. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\dependencies" -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.
  264. 6\include" --keep-dir x64\RelWithDebInfo -maxrregcount=0 --machine 64 --compile -cudart static --extended-lambd
  265. a --expt-relaxed-constexpr -std=c++14 -Xcompiler="/EHsc -Zi -Ob1 -bigobj" -D_WINDOWS -DNDEBUG -DNGP_GUI -DNGP_OPTIX
  266. -DTCNN_MIN_GPU_ARCH=75 -DTCNN_SHAMPOO -D"CMAKE_INTDIR=\"RelWithDebInfo\"" -D_MBCS -D"CMAKE_INTDIR=\"RelWithDebInfo\"
  267. " -Xcompiler "/EHsc /W1 /nologo /O2 /FdC:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\RelWithDebInfo\tiny-cud
  268. a-nn.pdb /FS /Zi /MD /GR" -o tiny-cuda-nn.dir\RelWithDebInfo\object.obj "C:\ngp\instant-ngp\dependencies\tiny-cuda-n
  269. n\src\object.cu"
  270.  
  271. C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src>"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\bin\
  272. nvcc.exe" -gencode=arch=compute_52,code=\"sm_52,compute_52\" --use-local-env -ccbin "C:\Program Files (x86)\Microsoft
  273. Visual Studio\2019\Community\VC\Tools\MSVC\14.29.30133\bin\HostX64\x64" -x cu -I"C:\ngp\instant-ngp\dependencies"
  274. -I"C:\ngp\instant-ngp\dependencies\eigen" -I"C:\ngp\instant-ngp\dependencies\filesystem" -I"C:\ngp\instant-ngp\depend
  275. encies\glfw\include" -I"C:\ngp\instant-ngp\dependencies\imgui\gl3w" -I"C:\ngp\instant-ngp\dependencies\nanovdb" -I"C:
  276. \ProgramData\NVIDIA Corporation\OptiX SDK 7.4.0\include" -I"C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include" -I"
  277. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\dependencies" -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.
  278. 6\include" --keep-dir x64\RelWithDebInfo -maxrregcount=0 --machine 64 --compile -cudart static --extended-lambd
  279. a --expt-relaxed-constexpr -std=c++14 -Xcompiler="/EHsc -Zi -Ob1 -bigobj" -D_WINDOWS -DNDEBUG -DNGP_GUI -DNGP_OPTIX
  280. -DTCNN_MIN_GPU_ARCH=75 -DTCNN_SHAMPOO -D"CMAKE_INTDIR=\"RelWithDebInfo\"" -D_MBCS -D"CMAKE_INTDIR=\"RelWithDebInfo\"
  281. " -Xcompiler "/EHsc /W1 /nologo /O2 /FdC:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\RelWithDebInfo\tiny-cud
  282. a-nn.pdb /FS /Zi /MD /GR" -o tiny-cuda-nn.dir\RelWithDebInfo\encoding.obj "C:\ngp\instant-ngp\dependencies\tiny-cuda
  283. -nn\src\encoding.cu"
  284.  
  285. C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src>"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\bin\
  286. nvcc.exe" -gencode=arch=compute_52,code=\"sm_52,compute_52\" --use-local-env -ccbin "C:\Program Files (x86)\Microsoft
  287. Visual Studio\2019\Community\VC\Tools\MSVC\14.29.30133\bin\HostX64\x64" -x cu -I"C:\ngp\instant-ngp\dependencies"
  288. -I"C:\ngp\instant-ngp\dependencies\eigen" -I"C:\ngp\instant-ngp\dependencies\filesystem" -I"C:\ngp\instant-ngp\depend
  289. encies\glfw\include" -I"C:\ngp\instant-ngp\dependencies\imgui\gl3w" -I"C:\ngp\instant-ngp\dependencies\nanovdb" -I"C:
  290. \ProgramData\NVIDIA Corporation\OptiX SDK 7.4.0\include" -I"C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include" -I"
  291. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\dependencies" -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.
  292. 6\include" --keep-dir x64\RelWithDebInfo -maxrregcount=0 --machine 64 --compile -cudart static --extended-lambd
  293. a --expt-relaxed-constexpr -std=c++14 -Xcompiler="/EHsc -Zi -Ob1 -bigobj" -D_WINDOWS -DNDEBUG -DNGP_GUI -DNGP_OPTIX
  294. -DTCNN_MIN_GPU_ARCH=75 -DTCNN_SHAMPOO -D"CMAKE_INTDIR=\"RelWithDebInfo\"" -D_MBCS -D"CMAKE_INTDIR=\"RelWithDebInfo\"
  295. " -Xcompiler "/EHsc /W1 /nologo /O2 /FdC:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\RelWithDebInfo\tiny-cud
  296. a-nn.pdb /FS /Zi /MD /GR" -o tiny-cuda-nn.dir\RelWithDebInfo\fully_fused_mlp.obj "C:\ngp\instant-ngp\dependencies\ti
  297. ny-cuda-nn\src\fully_fused_mlp.cu"
  298. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(415): error : name followed by "::" must be a class
  299. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  300.  
  301. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(493): error : name followed by "::" must be a class
  302. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  303.  
  304. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(493): error : name followed by "::" must be a class
  305. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  306.  
  307. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/common_device.h(74): error : more than one conversion
  308. function from "tcnn::network_precision_t" to a built-in type applies: [C:\ngp\instant-ngp\build\dependencies\tiny-cuda
  309. -nn\src\tiny-cuda-nn.vcxproj]
  310. function "__half::operator float() const"
  311. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(204): here
  312. function "__half::operator short() const"
  313. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(222): here
  314. function "__half::operator unsigned short() const"
  315. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(225): here
  316. function "__half::operator int() const"
  317. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(228): here
  318. function "__half::operator unsigned int() const"
  319. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(231): here
  320. function "__half::operator long long() const"
  321. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(234): here
  322. function "__half::operator unsigned long long() const"
  323. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(237): here
  324. function "__half::operator __nv_bool() const"
  325. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(241): here
  326. detected during:
  327. instantiation of "void tcnn::warp_activation<T,fragment_t>(tcnn::Activation, const fragment_t &, fragment
  328. _t &) [with T=tcnn::network_precision_t, fragment_t=tcnn::VectorFragment<tcnn::vector_t<tcnn::network_precision_t, 8U
  329. >>]"
  330. (244): here
  331. instantiation of "void tcnn::kernel_activation(uint32_t, tcnn::Activation, const T *, T *) [with T=tcnn::
  332. network_precision_t, N=8U]"
  333. (286): here
  334. instantiation of "void tcnn::activation_gpu(cudaStream_t, uint32_t, tcnn::Activation, const T *, T *) [wi
  335. th T=tcnn::network_precision_t]"
  336. (295): here
  337. instantiation of "void tcnn::activation_gpu(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrixDynamic<
  338. T> &, tcnn::GPUMatrixDynamic<T> &) [with T=tcnn::network_precision_t]"
  339. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_mlp.cu(149): here
  340. instantiation of "__nv_bool tcnn::compute_layer<CutlassLayer,T>(cudaStream_t, __nv_bool, tcnn::Activation
  341. , const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic
  342. <T> &, tcnn::GPUMatrixDynamic<T> &) [with CutlassLayer=tcnn::LastLayer, T=tcnn::network_precision_t]"
  343. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_mlp.cu(163): here
  344. instantiation of "__nv_bool tcnn::compute_inference_layer<CutlassLayer,T>(cudaStream_t, tcnn::Activation,
  345. const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<
  346. T> &) [with CutlassLayer=tcnn::LastLayer, T=tcnn::network_precision_t]"
  347. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_mlp.cu(183): here
  348. instantiation of "void tcnn::CutlassMLP<T>::inference_mixed_precision(cudaStream_t, const tcnn::GPUMatrix
  349. Dynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t]"
  350. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_mlp.cu(117): here
  351.  
  352. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/common_device.h(74): error : more than one conversion
  353. function from "tcnn::network_precision_t" to a built-in type applies: [C:\ngp\instant-ngp\build\dependencies\tiny-cuda
  354. -nn\src\tiny-cuda-nn.vcxproj]
  355. function "__half::operator float() const"
  356. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(204): here
  357. function "__half::operator short() const"
  358. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(222): here
  359. function "__half::operator unsigned short() const"
  360. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(225): here
  361. function "__half::operator int() const"
  362. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(228): here
  363. function "__half::operator unsigned int() const"
  364. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(231): here
  365. function "__half::operator long long() const"
  366. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(234): here
  367. function "__half::operator unsigned long long() const"
  368. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(237): here
  369. function "__half::operator __nv_bool() const"
  370. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(241): here
  371. detected during:
  372. instantiation of "void tcnn::warp_activation<T,fragment_t>(tcnn::Activation, const fragment_t &, fragment
  373. _t &) [with T=tcnn::network_precision_t, fragment_t=tcnn::VectorFragment<tcnn::vector_t<tcnn::network_precision_t, 8U
  374. >>]"
  375. (244): here
  376. instantiation of "void tcnn::kernel_activation(uint32_t, tcnn::Activation, const T *, T *) [with T=tcnn::
  377. network_precision_t, N=8U]"
  378. (286): here
  379. instantiation of "void tcnn::activation_gpu(cudaStream_t, uint32_t, tcnn::Activation, const T *, T *) [wi
  380. th T=tcnn::network_precision_t]"
  381. (295): here
  382. instantiation of "void tcnn::activation_gpu(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrixDynamic<
  383. T> &, tcnn::GPUMatrixDynamic<T> &) [with T=tcnn::network_precision_t]"
  384. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_mlp.cu(149): here
  385. instantiation of "__nv_bool tcnn::compute_layer<CutlassLayer,T>(cudaStream_t, __nv_bool, tcnn::Activation
  386. , const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic
  387. <T> &, tcnn::GPUMatrixDynamic<T> &) [with CutlassLayer=tcnn::LastLayer, T=tcnn::network_precision_t]"
  388. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_mlp.cu(163): here
  389. instantiation of "__nv_bool tcnn::compute_inference_layer<CutlassLayer,T>(cudaStream_t, tcnn::Activation,
  390. const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<
  391. T> &) [with CutlassLayer=tcnn::LastLayer, T=tcnn::network_precision_t]"
  392. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_mlp.cu(183): here
  393. instantiation of "void tcnn::CutlassMLP<T>::inference_mixed_precision(cudaStream_t, const tcnn::GPUMatrix
  394. Dynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t]"
  395. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_mlp.cu(117): here
  396.  
  397. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/encodings/grid.h(249): error : no operator "+=" match
  398. es these operands [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  399. operand types are: __half += __half
  400. detected during:
  401. instantiation of "void tcnn::kernel_grid<T,N_POS_DIMS,N_FEATURES_PER_LEVEL>(uint32_t, uint32_t, const uin
  402. t32_t *, uint32_t, float, float, float, const float *, tcnn::InterpolationType, tcnn::GridType, const T *, const floa
  403. t *, T *, float *) [with T=__half, N_POS_DIMS=2U, N_FEATURES_PER_LEVEL=1U]"
  404. (679): here
  405. instantiation of "void tcnn::GridEncodingTemplated<T, N_POS_DIMS, N_FEATURES_PER_LEVEL>::encode(cudaStrea
  406. m_t, uint32_t, tcnn::PitchedPtr<const float>, tcnn::PitchedPtr<T>, float *, __nv_bool) const [with T=__half, N_POS_DI
  407. MS=2U, N_FEATURES_PER_LEVEL=1U]"
  408. (608): here
  409. implicit generation of "tcnn::GridEncodingTemplated<T, N_POS_DIMS, N_FEATURES_PER_LEVEL>::~GridEncodingTe
  410. mplated() [with T=__half, N_POS_DIMS=2U, N_FEATURES_PER_LEVEL=1U]"
  411. (608): here
  412. instantiation of class "tcnn::GridEncodingTemplated<T, N_POS_DIMS, N_FEATURES_PER_LEVEL> [with T=__half,
  413. N_POS_DIMS=2U, N_FEATURES_PER_LEVEL=1U]"
  414. (608): here
  415. instantiation of "tcnn::GridEncodingTemplated<T, N_POS_DIMS, N_FEATURES_PER_LEVEL>::GridEncodingTemplated
  416. (uint32_t, uint32_t, uint32_t, float, __nv_bool, tcnn::InterpolationType, tcnn::GridType) [with T=__half, N_POS_DIMS=
  417. 2U, N_FEATURES_PER_LEVEL=1U]"
  418. (932): here
  419. instantiation of "tcnn::GridEncoding<T> *tcnn::create_grid_encoding_templated<T,N_FEATURES_PER_LEVEL>(uin
  420. t32_t, const tcnn::json &) [with T=__half, N_FEATURES_PER_LEVEL=1U]"
  421. (944): here
  422. instantiation of "tcnn::GridEncoding<T> *tcnn::create_grid_encoding<T>(uint32_t, const tcnn::json &) [wit
  423. h T=__half]"
  424. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\encoding.cu(126): here
  425. instantiation of "tcnn::Encoding<T> *tcnn::create_encoding<T>(uint32_t, const tcnn::json &, uint32_t) [wi
  426. th T=__half]"
  427. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/encodings/composite.h(84): here
  428.  
  429. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/common_device.h(187): error : more than one conversio
  430. n function from "tcnn::network_precision_t" to a built-in type applies: [C:\ngp\instant-ngp\build\dependencies\tiny-cud
  431. a-nn\src\tiny-cuda-nn.vcxproj]
  432. function "__half::operator float() const"
  433. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(204): here
  434. function "__half::operator short() const"
  435. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(222): here
  436. function "__half::operator unsigned short() const"
  437. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(225): here
  438. function "__half::operator int() const"
  439. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(228): here
  440. function "__half::operator unsigned int() const"
  441. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(231): here
  442. function "__half::operator long long() const"
  443. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(234): here
  444. function "__half::operator unsigned long long() const"
  445. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(237): here
  446. function "__half::operator __nv_bool() const"
  447. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(241): here
  448. detected during:
  449. instantiation of "void tcnn::warp_activation_backward<T,fragment_t,forward_fragment_t>(tcnn::Activation,
  450. const fragment_t &, const forward_fragment_t &, fragment_t &) [with T=tcnn::network_precision_t, fragment_t=tcnn::Vec
  451. torFragment<tcnn::vector_t<tcnn::network_precision_t, 8U>>, forward_fragment_t=tcnn::VectorFragment<tcnn::vector_t<tc
  452. nn::network_precision_t, 8U>>]"
  453. (269): here
  454. instantiation of "void tcnn::kernel_activation_backward_output(uint32_t, tcnn::Activation, const T *, con
  455. st T *, T *) [with T=tcnn::network_precision_t, N=8U]"
  456. (334): here
  457. instantiation of "void tcnn::activation_backward_output_gpu(cudaStream_t, uint32_t, tcnn::Activation, con
  458. st T *, const T *, T *) [with T=tcnn::network_precision_t]"
  459. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_mlp.cu(300): here
  460. instantiation of "void tcnn::CutlassMLP<T>::backward(cudaStream_t, const tcnn::Context &, const tcnn::GPU
  461. MatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> *
  462. , __nv_bool, __nv_bool) [with T=tcnn::network_precision_t]"
  463. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_mlp.cu(452): here
  464.  
  465. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/common_device.h(187): error : more than one conversio
  466. n function from "tcnn::network_precision_t" to a built-in type applies: [C:\ngp\instant-ngp\build\dependencies\tiny-cud
  467. a-nn\src\tiny-cuda-nn.vcxproj]
  468. function "__half::operator float() const"
  469. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(204): here
  470. function "__half::operator short() const"
  471. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(222): here
  472. function "__half::operator unsigned short() const"
  473. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(225): here
  474. function "__half::operator int() const"
  475. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(228): here
  476. function "__half::operator unsigned int() const"
  477. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(231): here
  478. function "__half::operator long long() const"
  479. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(234): here
  480. function "__half::operator unsigned long long() const"
  481. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(237): here
  482. function "__half::operator __nv_bool() const"
  483. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(241): here
  484. detected during:
  485. instantiation of "void tcnn::warp_activation_backward<T,fragment_t,forward_fragment_t>(tcnn::Activation,
  486. const fragment_t &, const forward_fragment_t &, fragment_t &) [with T=tcnn::network_precision_t, fragment_t=tcnn::Vec
  487. torFragment<tcnn::vector_t<tcnn::network_precision_t, 8U>>, forward_fragment_t=tcnn::VectorFragment<tcnn::vector_t<tc
  488. nn::network_precision_t, 8U>>]"
  489. (269): here
  490. instantiation of "void tcnn::kernel_activation_backward_output(uint32_t, tcnn::Activation, const T *, con
  491. st T *, T *) [with T=tcnn::network_precision_t, N=8U]"
  492. (334): here
  493. instantiation of "void tcnn::activation_backward_output_gpu(cudaStream_t, uint32_t, tcnn::Activation, con
  494. st T *, const T *, T *) [with T=tcnn::network_precision_t]"
  495. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_mlp.cu(300): here
  496. instantiation of "void tcnn::CutlassMLP<T>::backward(cudaStream_t, const tcnn::Context &, const tcnn::GPU
  497. MatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> *
  498. , __nv_bool, __nv_bool) [with T=tcnn::network_precision_t]"
  499. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_mlp.cu(452): here
  500.  
  501. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/common_device.h(193): error : more than one conversio
  502. n function from "tcnn::network_precision_t" to a built-in type applies: [C:\ngp\instant-ngp\build\dependencies\tiny-cud
  503. a-nn\src\tiny-cuda-nn.vcxproj]
  504. function "__half::operator float() const"
  505. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(204): here
  506. function "__half::operator short() const"
  507. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(222): here
  508. function "__half::operator unsigned short() const"
  509. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(225): here
  510. function "__half::operator int() const"
  511. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(228): here
  512. function "__half::operator unsigned int() const"
  513. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(231): here
  514. function "__half::operator long long() const"
  515. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(234): here
  516. function "__half::operator unsigned long long() const"
  517. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(237): here
  518. function "__half::operator __nv_bool() const"
  519. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(241): here
  520. detected during:
  521. instantiation of "void tcnn::warp_activation_backward<T,fragment_t,forward_fragment_t>(tcnn::Activation,
  522. const fragment_t &, const forward_fragment_t &, fragment_t &) [with T=tcnn::network_precision_t, fragment_t=tcnn::Vec
  523. torFragment<tcnn::vector_t<tcnn::network_precision_t, 8U>>, forward_fragment_t=tcnn::VectorFragment<tcnn::vector_t<tc
  524. nn::network_precision_t, 8U>>]"
  525. (269): here
  526. instantiation of "void tcnn::kernel_activation_backward_output(uint32_t, tcnn::Activation, const T *, con
  527. st T *, T *) [with T=tcnn::network_precision_t, N=8U]"
  528. (334): here
  529. instantiation of "void tcnn::activation_backward_output_gpu(cudaStream_t, uint32_t, tcnn::Activation, con
  530. st T *, const T *, T *) [with T=tcnn::network_precision_t]"
  531. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_mlp.cu(300): here
  532. instantiation of "void tcnn::CutlassMLP<T>::backward(cudaStream_t, const tcnn::Context &, const tcnn::GPU
  533. MatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> *
  534. , __nv_bool, __nv_bool) [with T=tcnn::network_precision_t]"
  535. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_mlp.cu(452): here
  536.  
  537. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/encodings/grid.h(249): error : no operator "+=" match
  538. es these operands [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  539. operand types are: __half += __half
  540. detected during:
  541. instantiation of "void tcnn::kernel_grid<T,N_POS_DIMS,N_FEATURES_PER_LEVEL>(uint32_t, uint32_t, const uin
  542. t32_t *, uint32_t, float, float, float, const float *, tcnn::InterpolationType, tcnn::GridType, const T *, const floa
  543. t *, T *, float *) [with T=__half, N_POS_DIMS=3U, N_FEATURES_PER_LEVEL=1U]"
  544. (679): here
  545. instantiation of "void tcnn::GridEncodingTemplated<T, N_POS_DIMS, N_FEATURES_PER_LEVEL>::encode(cudaStrea
  546. m_t, uint32_t, tcnn::PitchedPtr<const float>, tcnn::PitchedPtr<T>, float *, __nv_bool) const [with T=__half, N_POS_DI
  547. MS=3U, N_FEATURES_PER_LEVEL=1U]"
  548. (608): here
  549. implicit generation of "tcnn::GridEncodingTemplated<T, N_POS_DIMS, N_FEATURES_PER_LEVEL>::~GridEncodingTe
  550. mplated() [with T=__half, N_POS_DIMS=3U, N_FEATURES_PER_LEVEL=1U]"
  551. (608): here
  552. instantiation of class "tcnn::GridEncodingTemplated<T, N_POS_DIMS, N_FEATURES_PER_LEVEL> [with T=__half,
  553. N_POS_DIMS=3U, N_FEATURES_PER_LEVEL=1U]"
  554. (608): here
  555. instantiation of "tcnn::GridEncodingTemplated<T, N_POS_DIMS, N_FEATURES_PER_LEVEL>::GridEncodingTemplated
  556. (uint32_t, uint32_t, uint32_t, float, __nv_bool, tcnn::InterpolationType, tcnn::GridType) [with T=__half, N_POS_DIMS=
  557. 3U, N_FEATURES_PER_LEVEL=1U]"
  558. (933): here
  559. instantiation of "tcnn::GridEncoding<T> *tcnn::create_grid_encoding_templated<T,N_FEATURES_PER_LEVEL>(uin
  560. t32_t, const tcnn::json &) [with T=__half, N_FEATURES_PER_LEVEL=1U]"
  561. (944): here
  562. instantiation of "tcnn::GridEncoding<T> *tcnn::create_grid_encoding<T>(uint32_t, const tcnn::json &) [wit
  563. h T=__half]"
  564. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\encoding.cu(126): here
  565. instantiation of "tcnn::Encoding<T> *tcnn::create_encoding<T>(uint32_t, const tcnn::json &, uint32_t) [wi
  566. th T=__half]"
  567. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/encodings/composite.h(84): here
  568. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/common_device.h(193): error : more than one conversio
  569. n function from "tcnn::network_precision_t" to a built-in type applies: [C:\ngp\instant-ngp\build\dependencies\tiny-cud
  570. a-nn\src\tiny-cuda-nn.vcxproj]
  571. function "__half::operator float() const"
  572. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(204): here
  573. function "__half::operator short() const"
  574. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(222): here
  575. function "__half::operator unsigned short() const"
  576. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(225): here
  577. function "__half::operator int() const"
  578. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(228): here
  579. function "__half::operator unsigned int() const"
  580. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(231): here
  581. function "__half::operator long long() const"
  582. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(234): here
  583. function "__half::operator unsigned long long() const"
  584. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(237): here
  585. function "__half::operator __nv_bool() const"
  586. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(241): here
  587. detected during:
  588. instantiation of "void tcnn::warp_activation_backward<T,fragment_t,forward_fragment_t>(tcnn::Activation,
  589. const fragment_t &, const forward_fragment_t &, fragment_t &) [with T=tcnn::network_precision_t, fragment_t=tcnn::Vec
  590. torFragment<tcnn::vector_t<tcnn::network_precision_t, 8U>>, forward_fragment_t=tcnn::VectorFragment<tcnn::vector_t<tc
  591. nn::network_precision_t, 8U>>]"
  592. (269): here
  593. instantiation of "void tcnn::kernel_activation_backward_output(uint32_t, tcnn::Activation, const T *, con
  594. st T *, T *) [with T=tcnn::network_precision_t, N=8U]"
  595. (334): here
  596. instantiation of "void tcnn::activation_backward_output_gpu(cudaStream_t, uint32_t, tcnn::Activation, con
  597. st T *, const T *, T *) [with T=tcnn::network_precision_t]"
  598. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_mlp.cu(300): here
  599. instantiation of "void tcnn::CutlassMLP<T>::backward(cudaStream_t, const tcnn::Context &, const tcnn::GPU
  600. MatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> *
  601. , __nv_bool, __nv_bool) [with T=tcnn::network_precision_t]"
  602. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_mlp.cu(452): here
  603.  
  604.  
  605. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/common_device.h(204): error : more than one conversio
  606. n function from "tcnn::network_precision_t" to a built-in type applies: [C:\ngp\instant-ngp\build\dependencies\tiny-cud
  607. a-nn\src\tiny-cuda-nn.vcxproj]
  608. function "__half::operator float() const"
  609. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(204): here
  610. function "__half::operator short() const"
  611. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(222): here
  612. function "__half::operator unsigned short() const"
  613. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(225): here
  614. function "__half::operator int() const"
  615. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(228): here
  616. function "__half::operator unsigned int() const"
  617. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(231): here
  618. function "__half::operator long long() const"
  619. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(234): here
  620. function "__half::operator unsigned long long() const"
  621. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(237): here
  622. function "__half::operator __nv_bool() const"
  623. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(241): here
  624. detected during:
  625. instantiation of "void tcnn::warp_activation_backward<T,fragment_t,forward_fragment_t>(tcnn::Activation,
  626. const fragment_t &, const forward_fragment_t &, fragment_t &) [with T=tcnn::network_precision_t, fragment_t=tcnn::Vec
  627. torFragment<tcnn::vector_t<tcnn::network_precision_t, 8U>>, forward_fragment_t=tcnn::VectorFragment<tcnn::vector_t<tc
  628. nn::network_precision_t, 8U>>]"
  629. (269): here
  630. instantiation of "void tcnn::kernel_activation_backward_output(uint32_t, tcnn::Activation, const T *, con
  631. st T *, T *) [with T=tcnn::network_precision_t, N=8U]"
  632. (334): here
  633. instantiation of "void tcnn::activation_backward_output_gpu(cudaStream_t, uint32_t, tcnn::Activation, con
  634. st T *, const T *, T *) [with T=tcnn::network_precision_t]"
  635. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_mlp.cu(300): here
  636. instantiation of "void tcnn::CutlassMLP<T>::backward(cudaStream_t, const tcnn::Context &, const tcnn::GPU
  637. MatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> *
  638. , __nv_bool, __nv_bool) [with T=tcnn::network_precision_t]"
  639. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_mlp.cu(452): here
  640.  
  641. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/encodings/grid.h(249): error : no operator "+=" match
  642. es these operands [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  643. operand types are: __half += __half
  644. detected during:
  645. instantiation of "void tcnn::kernel_grid<T,N_POS_DIMS,N_FEATURES_PER_LEVEL>(uint32_t, uint32_t, const uin
  646. t32_t *, uint32_t, float, float, float, const float *, tcnn::InterpolationType, tcnn::GridType, const T *, const floa
  647. t *, T *, float *) [with T=__half, N_POS_DIMS=2U, N_FEATURES_PER_LEVEL=2U]"
  648. (679): here
  649. instantiation of "void tcnn::GridEncodingTemplated<T, N_POS_DIMS, N_FEATURES_PER_LEVEL>::encode(cudaStrea
  650. m_t, uint32_t, tcnn::PitchedPtr<const float>, tcnn::PitchedPtr<T>, float *, __nv_bool) const [with T=__half, N_POS_DI
  651. MS=2U, N_FEATURES_PER_LEVEL=2U]"
  652. (608): here
  653. implicit generation of "tcnn::GridEncodingTemplated<T, N_POS_DIMS, N_FEATURES_PER_LEVEL>::~GridEncodingTe
  654. mplated() [with T=__half, N_POS_DIMS=2U, N_FEATURES_PER_LEVEL=2U]"
  655. (608): here
  656. instantiation of class "tcnn::GridEncodingTemplated<T, N_POS_DIMS, N_FEATURES_PER_LEVEL> [with T=__half,
  657. N_POS_DIMS=2U, N_FEATURES_PER_LEVEL=2U]"
  658. (608): here
  659. instantiation of "tcnn::GridEncodingTemplated<T, N_POS_DIMS, N_FEATURES_PER_LEVEL>::GridEncodingTemplated
  660. (uint32_t, uint32_t, uint32_t, float, __nv_bool, tcnn::InterpolationType, tcnn::GridType) [with T=__half, N_POS_DIMS=
  661. 2U, N_FEATURES_PER_LEVEL=2U]"
  662. (932): here
  663. instantiation of "tcnn::GridEncoding<T> *tcnn::create_grid_encoding_templated<T,N_FEATURES_PER_LEVEL>(uin
  664. t32_t, const tcnn::json &) [with T=__half, N_FEATURES_PER_LEVEL=2U]"
  665. (945): here
  666. instantiation of "tcnn::GridEncoding<T> *tcnn::create_grid_encoding<T>(uint32_t, const tcnn::json &) [wit
  667. h T=__half]"
  668. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\encoding.cu(126): here
  669. instantiation of "tcnn::Encoding<T> *tcnn::create_encoding<T>(uint32_t, const tcnn::json &, uint32_t) [wi
  670. th T=__half]"
  671. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/encodings/composite.h(84): here
  672.  
  673. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/common_device.h(204): error : more than one conversio
  674. n function from "tcnn::network_precision_t" to a built-in type applies: [C:\ngp\instant-ngp\build\dependencies\tiny-cud
  675. a-nn\src\tiny-cuda-nn.vcxproj]
  676. function "__half::operator float() const"
  677. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(204): here
  678. function "__half::operator short() const"
  679. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(222): here
  680. function "__half::operator unsigned short() const"
  681. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(225): here
  682. function "__half::operator int() const"
  683. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(228): here
  684. function "__half::operator unsigned int() const"
  685. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(231): here
  686. function "__half::operator long long() const"
  687. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(234): here
  688. function "__half::operator unsigned long long() const"
  689. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(237): here
  690. function "__half::operator __nv_bool() const"
  691. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(241): here
  692. detected during:
  693. instantiation of "void tcnn::warp_activation_backward<T,fragment_t,forward_fragment_t>(tcnn::Activation,
  694. const fragment_t &, const forward_fragment_t &, fragment_t &) [with T=tcnn::network_precision_t, fragment_t=tcnn::Vec
  695. torFragment<tcnn::vector_t<tcnn::network_precision_t, 8U>>, forward_fragment_t=tcnn::VectorFragment<tcnn::vector_t<tc
  696. nn::network_precision_t, 8U>>]"
  697. (269): here
  698. instantiation of "void tcnn::kernel_activation_backward_output(uint32_t, tcnn::Activation, const T *, con
  699. st T *, T *) [with T=tcnn::network_precision_t, N=8U]"
  700. (334): here
  701. instantiation of "void tcnn::activation_backward_output_gpu(cudaStream_t, uint32_t, tcnn::Activation, con
  702. st T *, const T *, T *) [with T=tcnn::network_precision_t]"
  703. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_mlp.cu(300): here
  704. instantiation of "void tcnn::CutlassMLP<T>::backward(cudaStream_t, const tcnn::Context &, const tcnn::GPU
  705. MatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> *
  706. , __nv_bool, __nv_bool) [with T=tcnn::network_precision_t]"
  707. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_mlp.cu(452): here
  708.  
  709. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/encodings/grid.h(249): error : no operator "+=" match
  710. es these operands [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  711. operand types are: __half += __half
  712. detected during:
  713. instantiation of "void tcnn::kernel_grid<T,N_POS_DIMS,N_FEATURES_PER_LEVEL>(uint32_t, uint32_t, const uin
  714. t32_t *, uint32_t, float, float, float, const float *, tcnn::InterpolationType, tcnn::GridType, const T *, const floa
  715. t *, T *, float *) [with T=__half, N_POS_DIMS=3U, N_FEATURES_PER_LEVEL=2U]"
  716. (679): here
  717. instantiation of "void tcnn::GridEncodingTemplated<T, N_POS_DIMS, N_FEATURES_PER_LEVEL>::encode(cudaStrea
  718. m_t, uint32_t, tcnn::PitchedPtr<const float>, tcnn::PitchedPtr<T>, float *, __nv_bool) const [with T=__half, N_POS_DI
  719. MS=3U, N_FEATURES_PER_LEVEL=2U]"
  720. (608): here
  721. implicit generation of "tcnn::GridEncodingTemplated<T, N_POS_DIMS, N_FEATURES_PER_LEVEL>::~GridEncodingTe
  722. mplated() [with T=__half, N_POS_DIMS=3U, N_FEATURES_PER_LEVEL=2U]"
  723. (608): here
  724. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/common_device.h(211): error : more than one conversio
  725. n function from "tcnn::network_precision_t" to a built-in type applies: [C:\ngp\instant-ngp\build\dependencies\tiny-cud
  726. a-nn\src\tiny-cuda-nn.vcxproj]
  727. function "__half::operator float() const"
  728. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(204): here
  729. function "__half::operator short() const"
  730. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(222): here
  731. function "__half::operator unsigned short() const"
  732. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(225): here
  733. function "__half::operator int() const"
  734. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(228): here
  735. function "__half::operator unsigned int() const"
  736. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(231): here
  737. function "__half::operator long long() const"
  738. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(234): here
  739. function "__half::operator unsigned long long() const"
  740. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(237): here
  741. function "__half::operator __nv_bool() const"
  742. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(241): here
  743. detected during:
  744. instantiation of "void tcnn::warp_activation_backward<T,fragment_t,forward_fragment_t>(tcnn::Activation,
  745. const fragment_t &, const forward_fragment_t &, fragment_t &) [with T=tcnn::network_precision_t, fragment_t=tcnn::Vec
  746. torFragment<tcnn::vector_t<tcnn::network_precision_t, 8U>>, forward_fragment_t=tcnn::VectorFragment<tcnn::vector_t<tc
  747. nn::network_precision_t, 8U>>]"
  748. (269): here
  749. instantiation of "void tcnn::kernel_activation_backward_output(uint32_t, tcnn::Activation, const T *, con
  750. st T *, T *) [with T=tcnn::network_precision_t, N=8U]"
  751. (334): here
  752. instantiation of "void tcnn::activation_backward_output_gpu(cudaStream_t, uint32_t, tcnn::Activation, con
  753. st T *, const T *, T *) [with T=tcnn::network_precision_t]"
  754. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_mlp.cu(300): here
  755. instantiation of "void tcnn::CutlassMLP<T>::backward(cudaStream_t, const tcnn::Context &, const tcnn::GPU
  756. MatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> *
  757. , __nv_bool, __nv_bool) [with T=tcnn::network_precision_t]"
  758. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_mlp.cu(452): here
  759.  
  760. instantiation of class "tcnn::GridEncodingTemplated<T, N_POS_DIMS, N_FEATURES_PER_LEVEL> [with T=__half,
  761. N_POS_DIMS=3U, N_FEATURES_PER_LEVEL=2U]"
  762. (608): here
  763. instantiation of "tcnn::GridEncodingTemplated<T, N_POS_DIMS, N_FEATURES_PER_LEVEL>::GridEncodingTemplated
  764. (uint32_t, uint32_t, uint32_t, float, __nv_bool, tcnn::InterpolationType, tcnn::GridType) [with T=__half, N_POS_DIMS=
  765. 3U, N_FEATURES_PER_LEVEL=2U]"
  766. (933): here
  767. instantiation of "tcnn::GridEncoding<T> *tcnn::create_grid_encoding_templated<T,N_FEATURES_PER_LEVEL>(uin
  768. t32_t, const tcnn::json &) [with T=__half, N_FEATURES_PER_LEVEL=2U]"
  769. (945): here
  770. instantiation of "tcnn::GridEncoding<T> *tcnn::create_grid_encoding<T>(uint32_t, const tcnn::json &) [wit
  771. h T=__half]"
  772. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\encoding.cu(126): here
  773. instantiation of "tcnn::Encoding<T> *tcnn::create_encoding<T>(uint32_t, const tcnn::json &, uint32_t) [wi
  774. th T=__half]"
  775. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/encodings/composite.h(84): here
  776.  
  777. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/encodings/grid.h(249): error : no operator "+=" match
  778. es these operands [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  779. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/common_device.h(211): error : more than one conversio
  780. n function from "tcnn::network_precision_t" to a built-in type applies: [C:\ngp\instant-ngp\build\dependencies\tiny-cud
  781. a-nn\src\tiny-cuda-nn.vcxproj]
  782. function "__half::operator float() const"
  783. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(204): here
  784. function "__half::operator short() const"
  785. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(222): here
  786. function "__half::operator unsigned short() const"
  787. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(225): here
  788. function "__half::operator int() const"
  789. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(228): here
  790. function "__half::operator unsigned int() const"
  791. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(231): here
  792. function "__half::operator long long() const"
  793. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(234): here
  794. function "__half::operator unsigned long long() const"
  795. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(237): here
  796. function "__half::operator __nv_bool() const"
  797. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(241): here
  798. detected during:
  799. instantiation of "void tcnn::warp_activation_backward<T,fragment_t,forward_fragment_t>(tcnn::Activation,
  800. const fragment_t &, const forward_fragment_t &, fragment_t &) [with T=tcnn::network_precision_t, fragment_t=tcnn::Vec
  801. torFragment<tcnn::vector_t<tcnn::network_precision_t, 8U>>, forward_fragment_t=tcnn::VectorFragment<tcnn::vector_t<tc
  802. nn::network_precision_t, 8U>>]"
  803. (269): here
  804. instantiation of "void tcnn::kernel_activation_backward_output(uint32_t, tcnn::Activation, const T *, con
  805. st T *, T *) [with T=tcnn::network_precision_t, N=8U]"
  806. (334): here
  807. instantiation of "void tcnn::activation_backward_output_gpu(cudaStream_t, uint32_t, tcnn::Activation, con
  808. st T *, const T *, T *) [with T=tcnn::network_precision_t]"
  809. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_mlp.cu(300): here
  810. instantiation of "void tcnn::CutlassMLP<T>::backward(cudaStream_t, const tcnn::Context &, const tcnn::GPU
  811. MatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> *
  812. , __nv_bool, __nv_bool) [with T=tcnn::network_precision_t]"
  813. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_mlp.cu(452): here
  814.  
  815. operand types are: __half += __half
  816. detected during:
  817. instantiation of "void tcnn::kernel_grid<T,N_POS_DIMS,N_FEATURES_PER_LEVEL>(uint32_t, uint32_t, const uin
  818. t32_t *, uint32_t, float, float, float, const float *, tcnn::InterpolationType, tcnn::GridType, const T *, const floa
  819. t *, T *, float *) [with T=__half, N_POS_DIMS=2U, N_FEATURES_PER_LEVEL=4U]"
  820. (679): here
  821. instantiation of "void tcnn::GridEncodingTemplated<T, N_POS_DIMS, N_FEATURES_PER_LEVEL>::encode(cudaStrea
  822. m_t, uint32_t, tcnn::PitchedPtr<const float>, tcnn::PitchedPtr<T>, float *, __nv_bool) const [with T=__half, N_POS_DI
  823. MS=2U, N_FEATURES_PER_LEVEL=4U]"
  824. (608): here
  825. implicit generation of "tcnn::GridEncodingTemplated<T, N_POS_DIMS, N_FEATURES_PER_LEVEL>::~GridEncodingTe
  826. mplated() [with T=__half, N_POS_DIMS=2U, N_FEATURES_PER_LEVEL=4U]"
  827. (608): here
  828. instantiation of class "tcnn::GridEncodingTemplated<T, N_POS_DIMS, N_FEATURES_PER_LEVEL> [with T=__half,
  829. N_POS_DIMS=2U, N_FEATURES_PER_LEVEL=4U]"
  830. (608): here
  831. instantiation of "tcnn::GridEncodingTemplated<T, N_POS_DIMS, N_FEATURES_PER_LEVEL>::GridEncodingTemplated
  832. (uint32_t, uint32_t, uint32_t, float, __nv_bool, tcnn::InterpolationType, tcnn::GridType) [with T=__half, N_POS_DIMS=
  833. 2U, N_FEATURES_PER_LEVEL=4U]"
  834. (932): here
  835. instantiation of "tcnn::GridEncoding<T> *tcnn::create_grid_encoding_templated<T,N_FEATURES_PER_LEVEL>(uin
  836. t32_t, const tcnn::json &) [with T=__half, N_FEATURES_PER_LEVEL=4U]"
  837. (946): here
  838. instantiation of "tcnn::GridEncoding<T> *tcnn::create_grid_encoding<T>(uint32_t, const tcnn::json &) [wit
  839. h T=__half]"
  840. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\encoding.cu(126): here
  841. instantiation of "tcnn::Encoding<T> *tcnn::create_encoding<T>(uint32_t, const tcnn::json &, uint32_t) [wi
  842. th T=__half]"
  843. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/encodings/composite.h(84): here
  844.  
  845. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/encodings/grid.h(249): error : no operator "+=" match
  846. es these operands [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  847. operand types are: __half += __half
  848. detected during:
  849. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/common_device.h(217): error : more than one conversio
  850. n function from "tcnn::network_precision_t" to a built-in type applies: [C:\ngp\instant-ngp\build\dependencies\tiny-cud
  851. a-nn\src\tiny-cuda-nn.vcxproj]
  852. function "__half::operator float() const"
  853. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(204): here
  854. function "__half::operator short() const"
  855. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(222): here
  856. function "__half::operator unsigned short() const"
  857. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(225): here
  858. function "__half::operator int() const"
  859. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(228): here
  860. function "__half::operator unsigned int() const"
  861. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(231): here
  862. function "__half::operator long long() const"
  863. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(234): here
  864. function "__half::operator unsigned long long() const"
  865. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(237): here
  866. function "__half::operator __nv_bool() const"
  867. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(241): here
  868. detected during:
  869. instantiation of "void tcnn::warp_activation_backward<T,fragment_t,forward_fragment_t>(tcnn::Activation,
  870. const fragment_t &, const forward_fragment_t &, fragment_t &) [with T=tcnn::network_precision_t, fragment_t=tcnn::Vec
  871. torFragment<tcnn::vector_t<tcnn::network_precision_t, 8U>>, forward_fragment_t=tcnn::VectorFragment<tcnn::vector_t<tc
  872. nn::network_precision_t, 8U>>]"
  873. (269): here
  874. instantiation of "void tcnn::kernel_activation_backward_output(uint32_t, tcnn::Activation, const T *, con
  875. st T *, T *) [with T=tcnn::network_precision_t, N=8U]"
  876. (334): here
  877. instantiation of "void tcnn::activation_backward_output_gpu(cudaStream_t, uint32_t, tcnn::Activation, con
  878. st T *, const T *, T *) [with T=tcnn::network_precision_t]"
  879. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_mlp.cu(300): here
  880. instantiation of "void tcnn::CutlassMLP<T>::backward(cudaStream_t, const tcnn::Context &, const tcnn::GPU
  881. MatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> *
  882. , __nv_bool, __nv_bool) [with T=tcnn::network_precision_t]"
  883. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_mlp.cu(452): here
  884.  
  885. instantiation of "void tcnn::kernel_grid<T,N_POS_DIMS,N_FEATURES_PER_LEVEL>(uint32_t, uint32_t, const uin
  886. t32_t *, uint32_t, float, float, float, const float *, tcnn::InterpolationType, tcnn::GridType, const T *, const floa
  887. t *, T *, float *) [with T=__half, N_POS_DIMS=3U, N_FEATURES_PER_LEVEL=4U]"
  888. (679): here
  889. instantiation of "void tcnn::GridEncodingTemplated<T, N_POS_DIMS, N_FEATURES_PER_LEVEL>::encode(cudaStrea
  890. m_t, uint32_t, tcnn::PitchedPtr<const float>, tcnn::PitchedPtr<T>, float *, __nv_bool) const [with T=__half, N_POS_DI
  891. MS=3U, N_FEATURES_PER_LEVEL=4U]"
  892. (608): here
  893. implicit generation of "tcnn::GridEncodingTemplated<T, N_POS_DIMS, N_FEATURES_PER_LEVEL>::~GridEncodingTe
  894. mplated() [with T=__half, N_POS_DIMS=3U, N_FEATURES_PER_LEVEL=4U]"
  895. (608): here
  896. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/common_device.h(217): error : more than one conversio
  897. n function from "tcnn::network_precision_t" to a built-in type applies: [C:\ngp\instant-ngp\build\dependencies\tiny-cud
  898. a-nn\src\tiny-cuda-nn.vcxproj]
  899. function "__half::operator float() const"
  900. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(204): here
  901. function "__half::operator short() const"
  902. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(222): here
  903. function "__half::operator unsigned short() const"
  904. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(225): here
  905. function "__half::operator int() const"
  906. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(228): here
  907. function "__half::operator unsigned int() const"
  908. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(231): here
  909. function "__half::operator long long() const"
  910. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(234): here
  911. function "__half::operator unsigned long long() const"
  912. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(237): here
  913. function "__half::operator __nv_bool() const"
  914. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(241): here
  915. detected during:
  916. instantiation of "void tcnn::warp_activation_backward<T,fragment_t,forward_fragment_t>(tcnn::Activation,
  917. const fragment_t &, const forward_fragment_t &, fragment_t &) [with T=tcnn::network_precision_t, fragment_t=tcnn::Vec
  918. torFragment<tcnn::vector_t<tcnn::network_precision_t, 8U>>, forward_fragment_t=tcnn::VectorFragment<tcnn::vector_t<tc
  919. nn::network_precision_t, 8U>>]"
  920. (269): here
  921. instantiation of "void tcnn::kernel_activation_backward_output(uint32_t, tcnn::Activation, const T *, con
  922. st T *, T *) [with T=tcnn::network_precision_t, N=8U]"
  923. (334): here
  924. instantiation of "void tcnn::activation_backward_output_gpu(cudaStream_t, uint32_t, tcnn::Activation, con
  925. st T *, const T *, T *) [with T=tcnn::network_precision_t]"
  926. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_mlp.cu(300): here
  927. instantiation of "void tcnn::CutlassMLP<T>::backward(cudaStream_t, const tcnn::Context &, const tcnn::GPU
  928. MatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> *
  929. , __nv_bool, __nv_bool) [with T=tcnn::network_precision_t]"
  930. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_mlp.cu(452): here
  931.  
  932. instantiation of class "tcnn::GridEncodingTemplated<T, N_POS_DIMS, N_FEATURES_PER_LEVEL> [with T=__half,
  933. N_POS_DIMS=3U, N_FEATURES_PER_LEVEL=4U]"
  934. (608): here
  935. instantiation of "tcnn::GridEncodingTemplated<T, N_POS_DIMS, N_FEATURES_PER_LEVEL>::GridEncodingTemplated
  936. (uint32_t, uint32_t, uint32_t, float, __nv_bool, tcnn::InterpolationType, tcnn::GridType) [with T=__half, N_POS_DIMS=
  937. 3U, N_FEATURES_PER_LEVEL=4U]"
  938. (933): here
  939. instantiation of "tcnn::GridEncoding<T> *tcnn::create_grid_encoding_templated<T,N_FEATURES_PER_LEVEL>(uin
  940. t32_t, const tcnn::json &) [with T=__half, N_FEATURES_PER_LEVEL=4U]"
  941. (946): here
  942. instantiation of "tcnn::GridEncoding<T> *tcnn::create_grid_encoding<T>(uint32_t, const tcnn::json &) [wit
  943. h T=__half]"
  944. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\encoding.cu(126): here
  945. instantiation of "tcnn::Encoding<T> *tcnn::create_encoding<T>(uint32_t, const tcnn::json &, uint32_t) [wi
  946. th T=__half]"
  947. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/encodings/composite.h(84): here
  948.  
  949. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/encodings/grid.h(249): error : no operator "+=" match
  950. es these operands [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  951. operand types are: __half += __half
  952. detected during:
  953. instantiation of "void tcnn::kernel_grid<T,N_POS_DIMS,N_FEATURES_PER_LEVEL>(uint32_t, uint32_t, const uin
  954. t32_t *, uint32_t, float, float, float, const float *, tcnn::InterpolationType, tcnn::GridType, const T *, const floa
  955. t *, T *, float *) [with T=__half, N_POS_DIMS=2U, N_FEATURES_PER_LEVEL=8U]"
  956. (679): here
  957. instantiation of "void tcnn::GridEncodingTemplated<T, N_POS_DIMS, N_FEATURES_PER_LEVEL>::encode(cudaStrea
  958. m_t, uint32_t, tcnn::PitchedPtr<const float>, tcnn::PitchedPtr<T>, float *, __nv_bool) const [with T=__half, N_POS_DI
  959. MS=2U, N_FEATURES_PER_LEVEL=8U]"
  960. (608): here
  961. implicit generation of "tcnn::GridEncodingTemplated<T, N_POS_DIMS, N_FEATURES_PER_LEVEL>::~GridEncodingTe
  962. mplated() [with T=__half, N_POS_DIMS=2U, N_FEATURES_PER_LEVEL=8U]"
  963. (608): here
  964. instantiation of class "tcnn::GridEncodingTemplated<T, N_POS_DIMS, N_FEATURES_PER_LEVEL> [with T=__half,
  965. N_POS_DIMS=2U, N_FEATURES_PER_LEVEL=8U]"
  966. (608): here
  967. instantiation of "tcnn::GridEncodingTemplated<T, N_POS_DIMS, N_FEATURES_PER_LEVEL>::GridEncodingTemplated
  968. (uint32_t, uint32_t, uint32_t, float, __nv_bool, tcnn::InterpolationType, tcnn::GridType) [with T=__half, N_POS_DIMS=
  969. 2U, N_FEATURES_PER_LEVEL=8U]"
  970. (932): here
  971. instantiation of "tcnn::GridEncoding<T> *tcnn::create_grid_encoding_templated<T,N_FEATURES_PER_LEVEL>(uin
  972. t32_t, const tcnn::json &) [with T=__half, N_FEATURES_PER_LEVEL=8U]"
  973. (947): here
  974. instantiation of "tcnn::GridEncoding<T> *tcnn::create_grid_encoding<T>(uint32_t, const tcnn::json &) [wit
  975. h T=__half]"
  976. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\encoding.cu(126): here
  977. instantiation of "tcnn::Encoding<T> *tcnn::create_encoding<T>(uint32_t, const tcnn::json &, uint32_t) [wi
  978. th T=__half]"
  979. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/encodings/composite.h(84): here
  980.  
  981. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/encodings/grid.h(249): error : no operator "+=" match
  982. es these operands [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  983. operand types are: __half += __half
  984. detected during:
  985. instantiation of "void tcnn::kernel_grid<T,N_POS_DIMS,N_FEATURES_PER_LEVEL>(uint32_t, uint32_t, const uin
  986. t32_t *, uint32_t, float, float, float, const float *, tcnn::InterpolationType, tcnn::GridType, const T *, const floa
  987. t *, T *, float *) [with T=__half, N_POS_DIMS=3U, N_FEATURES_PER_LEVEL=8U]"
  988. (679): here
  989. instantiation of "void tcnn::GridEncodingTemplated<T, N_POS_DIMS, N_FEATURES_PER_LEVEL>::encode(cudaStrea
  990. m_t, uint32_t, tcnn::PitchedPtr<const float>, tcnn::PitchedPtr<T>, float *, __nv_bool) const [with T=__half, N_POS_DI
  991. MS=3U, N_FEATURES_PER_LEVEL=8U]"
  992. (608): here
  993. implicit generation of "tcnn::GridEncodingTemplated<T, N_POS_DIMS, N_FEATURES_PER_LEVEL>::~GridEncodingTe
  994. mplated() [with T=__half, N_POS_DIMS=3U, N_FEATURES_PER_LEVEL=8U]"
  995. (608): here
  996. instantiation of class "tcnn::GridEncodingTemplated<T, N_POS_DIMS, N_FEATURES_PER_LEVEL> [with T=__half,
  997. N_POS_DIMS=3U, N_FEATURES_PER_LEVEL=8U]"
  998. (608): here
  999. instantiation of "tcnn::GridEncodingTemplated<T, N_POS_DIMS, N_FEATURES_PER_LEVEL>::GridEncodingTemplated
  1000. (uint32_t, uint32_t, uint32_t, float, __nv_bool, tcnn::InterpolationType, tcnn::GridType) [with T=__half, N_POS_DIMS=
  1001. 3U, N_FEATURES_PER_LEVEL=8U]"
  1002. (933): here
  1003. instantiation of "tcnn::GridEncoding<T> *tcnn::create_grid_encoding_templated<T,N_FEATURES_PER_LEVEL>(uin
  1004. t32_t, const tcnn::json &) [with T=__half, N_FEATURES_PER_LEVEL=8U]"
  1005. (947): here
  1006. instantiation of "tcnn::GridEncoding<T> *tcnn::create_grid_encoding<T>(uint32_t, const tcnn::json &) [wit
  1007. h T=__half]"
  1008. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\encoding.cu(126): here
  1009. instantiation of "tcnn::Encoding<T> *tcnn::create_encoding<T>(uint32_t, const tcnn::json &, uint32_t) [wi
  1010. th T=__half]"
  1011. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/encodings/composite.h(84): here
  1012.  
  1013. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/common_device.h(519): error : more than one conversio
  1014. n function from "const tcnn::network_precision_t" to a built-in type applies: [C:\ngp\instant-ngp\build\dependencies\ti
  1015. ny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  1016. function "__half::operator float() const"
  1017. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(204): here
  1018. function "__half::operator short() const"
  1019. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(222): here
  1020. function "__half::operator unsigned short() const"
  1021. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(225): here
  1022. function "__half::operator int() const"
  1023. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(228): here
  1024. function "__half::operator unsigned int() const"
  1025. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(231): here
  1026. function "__half::operator long long() const"
  1027. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(234): here
  1028. function "__half::operator unsigned long long() const"
  1029. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(634): error : name followed by "::" must be a class
  1030. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  1031. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(237): here
  1032. function "__half::operator __nv_bool() const"
  1033. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(241): here
  1034. detected during:
  1035. instantiation of "void tcnn::add(uint32_t, const T *, T *) [with T=tcnn::network_precision_t]"
  1036. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_resnet.cu(156): here
  1037. instantiation of "void tcnn::CutlassResNet<T, input_activation>::inference_mixed_precision(cudaStream_t,
  1038. const tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, input_a
  1039. ctivation=tcnn::Activation::None]"
  1040. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_resnet.cu(106): here
  1041.  
  1042. detected during:
  1043. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/common_device.h(519): error : more than one conversio
  1044. n function from "tcnn::network_precision_t" to a built-in type applies: [C:\ngp\instant-ngp\build\dependencies\tiny-cud
  1045. a-nn\src\tiny-cuda-nn.vcxproj]
  1046. function "__half::operator float() const"
  1047. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(204): here
  1048. function "__half::operator short() const"
  1049. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(222): here
  1050. function "__half::operator unsigned short() const"
  1051. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(225): here
  1052. function "__half::operator int() const"
  1053. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(228): here
  1054. function "__half::operator unsigned int() const"
  1055. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(231): here
  1056. function "__half::operator long long() const"
  1057. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(234): here
  1058. function "__half::operator unsigned long long() const"
  1059. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(237): here
  1060. function "__half::operator __nv_bool() const"
  1061. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(241): here
  1062. detected during:
  1063. instantiation of "void tcnn::add(uint32_t, const T *, T *) [with T=tcnn::network_precision_t]"
  1064. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_resnet.cu(156): here
  1065. instantiation of "void tcnn::CutlassResNet<T, input_activation>::inference_mixed_precision(cudaStream_t,
  1066. const tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, input_a
  1067. ctivation=tcnn::Activation::None]"
  1068. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_resnet.cu(106): here
  1069.  
  1070. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  1071. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  1072. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  1073. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::None, INFERENCE=true]"
  1074. (760): here
  1075. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  1076. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  1077. (714): here
  1078.  
  1079. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(634): error : name followed by "::" must be a class
  1080. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  1081. detected during:
  1082. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  1083. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  1084. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  1085. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::None, INFERENCE=true]"
  1086. (760): here
  1087. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  1088. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  1089. (714): here
  1090.  
  1091. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(635): error : name followed by "::" must be a class
  1092. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  1093. detected during:
  1094. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  1095. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  1096. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  1097. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::None, INFERENCE=true]"
  1098. (760): here
  1099. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  1100. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  1101. (714): here
  1102.  
  1103. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(635): error : name followed by "::" must be a class
  1104. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  1105. detected during:
  1106. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  1107. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  1108. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  1109. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::None, INFERENCE=true]"
  1110. (760): here
  1111. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  1112. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  1113. (714): here
  1114.  
  1115. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(515): error : name followed by "::" must be a class
  1116. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  1117. detected during:
  1118. instantiation of "void tcnn::kernel_mlp_fused<WIDTH,N_ITERS,OUT_T,ACTIVATION,INFERENCE>(tcnn::Activation,
  1119. const __half *, const __half *, OUT_T *, OUT_T *, uint32_t, uint32_t, uint32_t, uint32_t, <error-type>, <error-type>
  1120. ) [with WIDTH=128, N_ITERS=8, OUT_T=__half, ACTIVATION=tcnn::Activation::None, INFERENCE=true]"
  1121. (636): here
  1122. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  1123. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  1124. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  1125. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::None, INFERENCE=true]"
  1126. (760): here
  1127. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/common_device.h(129): error : more than one conversio
  1128. n function from "tcnn::network_precision_t" to a built-in type applies: [C:\ngp\instant-ngp\build\dependencies\tiny-cud
  1129. a-nn\src\tiny-cuda-nn.vcxproj]
  1130. function "__half::operator float() const"
  1131. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(204): here
  1132. function "__half::operator short() const"
  1133. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(222): here
  1134. function "__half::operator unsigned short() const"
  1135. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(225): here
  1136. function "__half::operator int() const"
  1137. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(228): here
  1138. function "__half::operator unsigned int() const"
  1139. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(231): here
  1140. function "__half::operator long long() const"
  1141. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(234): here
  1142. function "__half::operator unsigned long long() const"
  1143. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(237): here
  1144. function "__half::operator __nv_bool() const"
  1145. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(241): here
  1146. detected during:
  1147. instantiation of "void tcnn::warp_activation_backward_in<T,fragment_t,forward_fragment_t>(tcnn::Activatio
  1148. n, const fragment_t &, const forward_fragment_t &, fragment_t &) [with T=tcnn::network_precision_t, fragment_t=tcnn::
  1149. VectorFragment<tcnn::vector_t<tcnn::network_precision_t, 8U>>, forward_fragment_t=tcnn::VectorFragment<tcnn::vector_t
  1150. <tcnn::network_precision_t, 8U>>]"
  1151. (256): here
  1152. instantiation of "void tcnn::kernel_activation_backward(uint32_t, tcnn::Activation, const T *, const T *,
  1153. T *) [with T=tcnn::network_precision_t, N=8U]"
  1154. (310): here
  1155. instantiation of "void tcnn::activation_backward_gpu(cudaStream_t, uint32_t, tcnn::Activation, const T *,
  1156. const T *, T *) [with T=tcnn::network_precision_t]"
  1157. (319): here
  1158. instantiation of "void tcnn::activation_backward_gpu(cudaStream_t, tcnn::Activation, const tcnn::GPUMatri
  1159. xDynamic<T> &, tcnn::GPUMatrixDynamic<T> &) [with T=tcnn::network_precision_t]"
  1160. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_mlp.cu(353): here
  1161. instantiation of "void tcnn::CutlassMLP<T>::backward(cudaStream_t, const tcnn::Context &, const tcnn::GPU
  1162. MatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> *
  1163. , __nv_bool, __nv_bool) [with T=tcnn::network_precision_t]"
  1164. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_mlp.cu(452): here
  1165.  
  1166. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  1167. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  1168. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/common_device.h(129): error : more than one conversio
  1169. n function from "tcnn::network_precision_t" to a built-in type applies: [C:\ngp\instant-ngp\build\dependencies\tiny-cud
  1170. a-nn\src\tiny-cuda-nn.vcxproj]
  1171. function "__half::operator float() const"
  1172. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(204): here
  1173. function "__half::operator short() const"
  1174. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(222): here
  1175. function "__half::operator unsigned short() const"
  1176. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(225): here
  1177. function "__half::operator int() const"
  1178. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(228): here
  1179. function "__half::operator unsigned int() const"
  1180. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(231): here
  1181. function "__half::operator long long() const"
  1182. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(234): here
  1183. function "__half::operator unsigned long long() const"
  1184. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(237): here
  1185. function "__half::operator __nv_bool() const"
  1186. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(241): here
  1187. detected during:
  1188. instantiation of "void tcnn::warp_activation_backward_in<T,fragment_t,forward_fragment_t>(tcnn::Activatio
  1189. n, const fragment_t &, const forward_fragment_t &, fragment_t &) [with T=tcnn::network_precision_t, fragment_t=tcnn::
  1190. VectorFragment<tcnn::vector_t<tcnn::network_precision_t, 8U>>, forward_fragment_t=tcnn::VectorFragment<tcnn::vector_t
  1191. <tcnn::network_precision_t, 8U>>]"
  1192. (256): here
  1193. instantiation of "void tcnn::kernel_activation_backward(uint32_t, tcnn::Activation, const T *, const T *,
  1194. T *) [with T=tcnn::network_precision_t, N=8U]"
  1195. (310): here
  1196. instantiation of "void tcnn::activation_backward_gpu(cudaStream_t, uint32_t, tcnn::Activation, const T *,
  1197. const T *, T *) [with T=tcnn::network_precision_t]"
  1198. (319): here
  1199. instantiation of "void tcnn::activation_backward_gpu(cudaStream_t, tcnn::Activation, const tcnn::GPUMatri
  1200. xDynamic<T> &, tcnn::GPUMatrixDynamic<T> &) [with T=tcnn::network_precision_t]"
  1201. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_mlp.cu(353): here
  1202. instantiation of "void tcnn::CutlassMLP<T>::backward(cudaStream_t, const tcnn::Context &, const tcnn::GPU
  1203. MatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> *
  1204. , __nv_bool, __nv_bool) [with T=tcnn::network_precision_t]"
  1205. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_mlp.cu(452): here
  1206.  
  1207. (714): here
  1208. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/common_device.h(135): error : more than one conversio
  1209. n function from "tcnn::network_precision_t" to a built-in type applies: [C:\ngp\instant-ngp\build\dependencies\tiny-cud
  1210. a-nn\src\tiny-cuda-nn.vcxproj]
  1211. function "__half::operator float() const"
  1212. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(204): here
  1213. function "__half::operator short() const"
  1214. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(222): here
  1215. function "__half::operator unsigned short() const"
  1216. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(225): here
  1217. function "__half::operator int() const"
  1218. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(228): here
  1219. function "__half::operator unsigned int() const"
  1220. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(231): here
  1221. function "__half::operator long long() const"
  1222. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(234): here
  1223. function "__half::operator unsigned long long() const"
  1224. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(237): here
  1225. function "__half::operator __nv_bool() const"
  1226. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(241): here
  1227. detected during:
  1228. instantiation of "void tcnn::warp_activation_backward_in<T,fragment_t,forward_fragment_t>(tcnn::Activatio
  1229. n, const fragment_t &, const forward_fragment_t &, fragment_t &) [with T=tcnn::network_precision_t, fragment_t=tcnn::
  1230. VectorFragment<tcnn::vector_t<tcnn::network_precision_t, 8U>>, forward_fragment_t=tcnn::VectorFragment<tcnn::vector_t
  1231. <tcnn::network_precision_t, 8U>>]"
  1232. (256): here
  1233. instantiation of "void tcnn::kernel_activation_backward(uint32_t, tcnn::Activation, const T *, const T *,
  1234. T *) [with T=tcnn::network_precision_t, N=8U]"
  1235. (310): here
  1236. instantiation of "void tcnn::activation_backward_gpu(cudaStream_t, uint32_t, tcnn::Activation, const T *,
  1237. const T *, T *) [with T=tcnn::network_precision_t]"
  1238. (319): here
  1239. instantiation of "void tcnn::activation_backward_gpu(cudaStream_t, tcnn::Activation, const tcnn::GPUMatri
  1240. xDynamic<T> &, tcnn::GPUMatrixDynamic<T> &) [with T=tcnn::network_precision_t]"
  1241. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_mlp.cu(353): here
  1242. instantiation of "void tcnn::CutlassMLP<T>::backward(cudaStream_t, const tcnn::Context &, const tcnn::GPU
  1243. MatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> *
  1244. , __nv_bool, __nv_bool) [with T=tcnn::network_precision_t]"
  1245. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_mlp.cu(452): here
  1246.  
  1247.  
  1248. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(516): error : name followed by "::" must be a class
  1249. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  1250. detected during:
  1251. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/common_device.h(135): error : more than one conversio
  1252. n function from "tcnn::network_precision_t" to a built-in type applies: [C:\ngp\instant-ngp\build\dependencies\tiny-cud
  1253. a-nn\src\tiny-cuda-nn.vcxproj]
  1254. function "__half::operator float() const"
  1255. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(204): here
  1256. function "__half::operator short() const"
  1257. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(222): here
  1258. function "__half::operator unsigned short() const"
  1259. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(225): here
  1260. function "__half::operator int() const"
  1261. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(228): here
  1262. function "__half::operator unsigned int() const"
  1263. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(231): here
  1264. function "__half::operator long long() const"
  1265. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(234): here
  1266. function "__half::operator unsigned long long() const"
  1267. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(237): here
  1268. function "__half::operator __nv_bool() const"
  1269. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(241): here
  1270. detected during:
  1271. instantiation of "void tcnn::warp_activation_backward_in<T,fragment_t,forward_fragment_t>(tcnn::Activatio
  1272. n, const fragment_t &, const forward_fragment_t &, fragment_t &) [with T=tcnn::network_precision_t, fragment_t=tcnn::
  1273. VectorFragment<tcnn::vector_t<tcnn::network_precision_t, 8U>>, forward_fragment_t=tcnn::VectorFragment<tcnn::vector_t
  1274. <tcnn::network_precision_t, 8U>>]"
  1275. (256): here
  1276. instantiation of "void tcnn::kernel_activation_backward(uint32_t, tcnn::Activation, const T *, const T *,
  1277. T *) [with T=tcnn::network_precision_t, N=8U]"
  1278. (310): here
  1279. instantiation of "void tcnn::activation_backward_gpu(cudaStream_t, uint32_t, tcnn::Activation, const T *,
  1280. const T *, T *) [with T=tcnn::network_precision_t]"
  1281. (319): here
  1282. instantiation of "void tcnn::activation_backward_gpu(cudaStream_t, tcnn::Activation, const tcnn::GPUMatri
  1283. xDynamic<T> &, tcnn::GPUMatrixDynamic<T> &) [with T=tcnn::network_precision_t]"
  1284. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_mlp.cu(353): here
  1285. instantiation of "void tcnn::CutlassMLP<T>::backward(cudaStream_t, const tcnn::Context &, const tcnn::GPU
  1286. MatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> *
  1287. , __nv_bool, __nv_bool) [with T=tcnn::network_precision_t]"
  1288. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_mlp.cu(452): here
  1289.  
  1290. instantiation of "void tcnn::kernel_mlp_fused<WIDTH,N_ITERS,OUT_T,ACTIVATION,INFERENCE>(tcnn::Activation,
  1291. const __half *, const __half *, OUT_T *, OUT_T *, uint32_t, uint32_t, uint32_t, uint32_t, <error-type>, <error-type>
  1292. ) [with WIDTH=128, N_ITERS=8, OUT_T=__half, ACTIVATION=tcnn::Activation::None, INFERENCE=true]"
  1293. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/common_device.h(141): error : more than one conversio
  1294. n function from "tcnn::network_precision_t" to a built-in type applies: [C:\ngp\instant-ngp\build\dependencies\tiny-cud
  1295. a-nn\src\tiny-cuda-nn.vcxproj]
  1296. function "__half::operator float() const"
  1297. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(204): here
  1298. function "__half::operator short() const"
  1299. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(222): here
  1300. function "__half::operator unsigned short() const"
  1301. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(225): here
  1302. function "__half::operator int() const"
  1303. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(228): here
  1304. function "__half::operator unsigned int() const"
  1305. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(231): here
  1306. function "__half::operator long long() const"
  1307. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(234): here
  1308. function "__half::operator unsigned long long() const"
  1309. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(237): here
  1310. function "__half::operator __nv_bool() const"
  1311. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(241): here
  1312. detected during:
  1313. instantiation of "void tcnn::warp_activation_backward_in<T,fragment_t,forward_fragment_t>(tcnn::Activatio
  1314. n, const fragment_t &, const forward_fragment_t &, fragment_t &) [with T=tcnn::network_precision_t, fragment_t=tcnn::
  1315. VectorFragment<tcnn::vector_t<tcnn::network_precision_t, 8U>>, forward_fragment_t=tcnn::VectorFragment<tcnn::vector_t
  1316. <tcnn::network_precision_t, 8U>>]"
  1317. (256): here
  1318. instantiation of "void tcnn::kernel_activation_backward(uint32_t, tcnn::Activation, const T *, const T *,
  1319. T *) [with T=tcnn::network_precision_t, N=8U]"
  1320. (310): here
  1321. instantiation of "void tcnn::activation_backward_gpu(cudaStream_t, uint32_t, tcnn::Activation, const T *,
  1322. const T *, T *) [with T=tcnn::network_precision_t]"
  1323. (319): here
  1324. instantiation of "void tcnn::activation_backward_gpu(cudaStream_t, tcnn::Activation, const tcnn::GPUMatri
  1325. xDynamic<T> &, tcnn::GPUMatrixDynamic<T> &) [with T=tcnn::network_precision_t]"
  1326. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_mlp.cu(353): here
  1327. instantiation of "void tcnn::CutlassMLP<T>::backward(cudaStream_t, const tcnn::Context &, const tcnn::GPU
  1328. MatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> *
  1329. , __nv_bool, __nv_bool) [with T=tcnn::network_precision_t]"
  1330. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_mlp.cu(452): here
  1331.  
  1332. (636): here
  1333. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/common_device.h(74): error : more than one conversion
  1334. function from "tcnn::network_precision_t" to a built-in type applies: [C:\ngp\instant-ngp\build\dependencies\tiny-cuda
  1335. -nn\src\tiny-cuda-nn.vcxproj]
  1336. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  1337. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  1338. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  1339. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::None, INFERENCE=true]"
  1340. (760): here
  1341. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  1342. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  1343. (714): here
  1344.  
  1345. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(517): error : name followed by "::" must be a class
  1346. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  1347. detected during:
  1348. instantiation of "void tcnn::kernel_mlp_fused<WIDTH,N_ITERS,OUT_T,ACTIVATION,INFERENCE>(tcnn::Activation,
  1349. const __half *, const __half *, OUT_T *, OUT_T *, uint32_t, uint32_t, uint32_t, uint32_t, <error-type>, <error-type>
  1350. ) [with WIDTH=128, N_ITERS=8, OUT_T=__half, ACTIVATION=tcnn::Activation::None, INFERENCE=true]"
  1351. (636): here
  1352. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  1353. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  1354. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  1355. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::None, INFERENCE=true]"
  1356. (760): here
  1357. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  1358. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  1359. (714): here
  1360.  
  1361. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(519): error : name followed by "::" must be a class
  1362. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  1363. detected during:
  1364. instantiation of "void tcnn::kernel_mlp_fused<WIDTH,N_ITERS,OUT_T,ACTIVATION,INFERENCE>(tcnn::Activation,
  1365. const __half *, const __half *, OUT_T *, OUT_T *, uint32_t, uint32_t, uint32_t, uint32_t, <error-type>, <error-type>
  1366. ) [with WIDTH=128, N_ITERS=8, OUT_T=__half, ACTIVATION=tcnn::Activation::None, INFERENCE=true]"
  1367. (636): here
  1368. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  1369. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  1370. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  1371. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::None, INFERENCE=true]"
  1372. (760): here
  1373. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  1374. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  1375. function "__half::operator float() const"
  1376. (714): here
  1377.  
  1378. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(59): error : name must be a namespace name [C:\ngp\
  1379. instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  1380. detected during:
  1381. instantiation of "void tcnn::threadblock_layer<WIDTH,N_ITERS,OUT_T,BACKWARD>(tcnn::Activation, __half *,
  1382. const __half *, OUT_T *, const OUT_T *) [with WIDTH=128, N_ITERS=8, OUT_T=__half, BACKWARD=false]"
  1383. (525): here
  1384. instantiation of "void tcnn::kernel_mlp_fused<WIDTH,N_ITERS,OUT_T,ACTIVATION,INFERENCE>(tcnn::Activation,
  1385. const __half *, const __half *, OUT_T *, OUT_T *, uint32_t, uint32_t, uint32_t, uint32_t, <error-type>, <error-type>
  1386. ) [with WIDTH=128, N_ITERS=8, OUT_T=__half, ACTIVATION=tcnn::Activation::None, INFERENCE=true]"
  1387. (636): here
  1388. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  1389. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  1390. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  1391. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::None, INFERENCE=true]"
  1392. (760): here
  1393. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  1394. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  1395. (714): here
  1396.  
  1397. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(63): error : name followed by "::" must be a class
  1398. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  1399. detected during:
  1400. instantiation of "void tcnn::threadblock_layer<WIDTH,N_ITERS,OUT_T,BACKWARD>(tcnn::Activation, __half *,
  1401. const __half *, OUT_T *, const OUT_T *) [with WIDTH=128, N_ITERS=8, OUT_T=__half, BACKWARD=false]"
  1402. (525): here
  1403. instantiation of "void tcnn::kernel_mlp_fused<WIDTH,N_ITERS,OUT_T,ACTIVATION,INFERENCE>(tcnn::Activation,
  1404. const __half *, const __half *, OUT_T *, OUT_T *, uint32_t, uint32_t, uint32_t, uint32_t, <error-type>, <error-type>
  1405. ) [with WIDTH=128, N_ITERS=8, OUT_T=__half, ACTIVATION=tcnn::Activation::None, INFERENCE=true]"
  1406. (636): here
  1407. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  1408. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  1409. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  1410. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::None, INFERENCE=true]"
  1411. (760): here
  1412. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  1413. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  1414. (714): here
  1415.  
  1416. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(63): error : name followed by "::" must be a class
  1417. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  1418. detected during:
  1419. instantiation of "void tcnn::threadblock_layer<WIDTH,N_ITERS,OUT_T,BACKWARD>(tcnn::Activation, __half *,
  1420. const __half *, OUT_T *, const OUT_T *) [with WIDTH=128, N_ITERS=8, OUT_T=__half, BACKWARD=false]"
  1421. (525): here
  1422. instantiation of "void tcnn::kernel_mlp_fused<WIDTH,N_ITERS,OUT_T,ACTIVATION,INFERENCE>(tcnn::Activation,
  1423. const __half *, const __half *, OUT_T *, OUT_T *, uint32_t, uint32_t, uint32_t, uint32_t, <error-type>, <error-type>
  1424. ) [with WIDTH=128, N_ITERS=8, OUT_T=__half, ACTIVATION=tcnn::Activation::None, INFERENCE=true]"
  1425. (636): here
  1426. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  1427. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  1428. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  1429. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::None, INFERENCE=true]"
  1430. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(204): here
  1431. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/common_device.h(141): error : more than one conversio
  1432. n function from "tcnn::network_precision_t" to a built-in type applies: [C:\ngp\instant-ngp\build\dependencies\tiny-cud
  1433. a-nn\src\tiny-cuda-nn.vcxproj]
  1434. function "__half::operator float() const"
  1435. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(204): here
  1436. function "__half::operator short() const"
  1437. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(222): here
  1438. function "__half::operator unsigned short() const"
  1439. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(225): here
  1440. function "__half::operator int() const"
  1441. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(228): here
  1442. function "__half::operator unsigned int() const"
  1443. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(231): here
  1444. function "__half::operator long long() const"
  1445. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(234): here
  1446. function "__half::operator unsigned long long() const"
  1447. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(237): here
  1448. function "__half::operator __nv_bool() const"
  1449. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(241): here
  1450. detected during:
  1451. instantiation of "void tcnn::warp_activation_backward_in<T,fragment_t,forward_fragment_t>(tcnn::Activatio
  1452. n, const fragment_t &, const forward_fragment_t &, fragment_t &) [with T=tcnn::network_precision_t, fragment_t=tcnn::
  1453. VectorFragment<tcnn::vector_t<tcnn::network_precision_t, 8U>>, forward_fragment_t=tcnn::VectorFragment<tcnn::vector_t
  1454. <tcnn::network_precision_t, 8U>>]"
  1455. (256): here
  1456. instantiation of "void tcnn::kernel_activation_backward(uint32_t, tcnn::Activation, const T *, const T *,
  1457. T *) [with T=tcnn::network_precision_t, N=8U]"
  1458. (310): here
  1459. instantiation of "void tcnn::activation_backward_gpu(cudaStream_t, uint32_t, tcnn::Activation, const T *,
  1460. const T *, T *) [with T=tcnn::network_precision_t]"
  1461. (319): here
  1462. instantiation of "void tcnn::activation_backward_gpu(cudaStream_t, tcnn::Activation, const tcnn::GPUMatri
  1463. xDynamic<T> &, tcnn::GPUMatrixDynamic<T> &) [with T=tcnn::network_precision_t]"
  1464. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_mlp.cu(353): here
  1465. instantiation of "void tcnn::CutlassMLP<T>::backward(cudaStream_t, const tcnn::Context &, const tcnn::GPU
  1466. MatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> *
  1467. , __nv_bool, __nv_bool) [with T=tcnn::network_precision_t]"
  1468. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_mlp.cu(452): here
  1469.  
  1470. (760): here
  1471. function "__half::operator short() const"
  1472. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(222): here
  1473. function "__half::operator unsigned short() const"
  1474. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(225): here
  1475. function "__half::operator int() const"
  1476. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(228): here
  1477. function "__half::operator unsigned int() const"
  1478. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(231): here
  1479. function "__half::operator long long() const"
  1480. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(234): here
  1481. function "__half::operator unsigned long long() const"
  1482. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(237): here
  1483. function "__half::operator __nv_bool() const"
  1484. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(241): here
  1485. detected during:
  1486. instantiation of "void tcnn::warp_activation<T,fragment_t>(tcnn::Activation, const fragment_t &, fragment
  1487. _t &) [with T=tcnn::network_precision_t, fragment_t=tcnn::VectorFragment<tcnn::vector_t<tcnn::network_precision_t, 8U
  1488. >>]"
  1489. (244): here
  1490. instantiation of "void tcnn::kernel_activation(uint32_t, tcnn::Activation, const T *, T *) [with T=tcnn::
  1491. network_precision_t, N=8U]"
  1492. (286): here
  1493. instantiation of "void tcnn::activation_gpu(cudaStream_t, uint32_t, tcnn::Activation, const T *, T *) [wi
  1494. th T=tcnn::network_precision_t]"
  1495. (295): here
  1496. instantiation of "void tcnn::activation_gpu(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrixDynamic<
  1497. T> &, tcnn::GPUMatrixDynamic<T> &) [with T=tcnn::network_precision_t]"
  1498. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_resnet.cu(202): here
  1499. instantiation of "std::unique_ptr<tcnn::Context, std::default_delete<tcnn::Context>> tcnn::CutlassResNet<
  1500. T, input_activation>::forward(cudaStream_t, const tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> *, __nv_bool
  1501. , __nv_bool) [with T=tcnn::network_precision_t, input_activation=tcnn::Activation::None]"
  1502. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_resnet.cu(391): here
  1503.  
  1504. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/common_device.h(74): error : more than one conversion
  1505. function from "tcnn::network_precision_t" to a built-in type applies: [C:\ngp\instant-ngp\build\dependencies\tiny-cuda
  1506. -nn\src\tiny-cuda-nn.vcxproj]
  1507. function "__half::operator float() const"
  1508. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(204): here
  1509. function "__half::operator short() const"
  1510. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(222): here
  1511. function "__half::operator unsigned short() const"
  1512. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(225): here
  1513. function "__half::operator int() const"
  1514. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(228): here
  1515. function "__half::operator unsigned int() const"
  1516. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(231): here
  1517. function "__half::operator long long() const"
  1518. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(234): here
  1519. function "__half::operator unsigned long long() const"
  1520. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(237): here
  1521. function "__half::operator __nv_bool() const"
  1522. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(241): here
  1523. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/common_device.h(148): error : more than one conversio
  1524. n function from "tcnn::network_precision_t" to a built-in type applies: [C:\ngp\instant-ngp\build\dependencies\tiny-cud
  1525. a-nn\src\tiny-cuda-nn.vcxproj]
  1526. function "__half::operator float() const"
  1527. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(204): here
  1528. function "__half::operator short() const"
  1529. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(222): here
  1530. function "__half::operator unsigned short() const"
  1531. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(225): here
  1532. function "__half::operator int() const"
  1533. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(228): here
  1534. function "__half::operator unsigned int() const"
  1535. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(231): here
  1536. function "__half::operator long long() const"
  1537. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(234): here
  1538. function "__half::operator unsigned long long() const"
  1539. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(237): here
  1540. function "__half::operator __nv_bool() const"
  1541. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(241): here
  1542. detected during:
  1543. instantiation of "void tcnn::warp_activation_backward_in<T,fragment_t,forward_fragment_t>(tcnn::Activatio
  1544. n, const fragment_t &, const forward_fragment_t &, fragment_t &) [with T=tcnn::network_precision_t, fragment_t=tcnn::
  1545. VectorFragment<tcnn::vector_t<tcnn::network_precision_t, 8U>>, forward_fragment_t=tcnn::VectorFragment<tcnn::vector_t
  1546. <tcnn::network_precision_t, 8U>>]"
  1547. (256): here
  1548. instantiation of "void tcnn::kernel_activation_backward(uint32_t, tcnn::Activation, const T *, const T *,
  1549. T *) [with T=tcnn::network_precision_t, N=8U]"
  1550. (310): here
  1551. instantiation of "void tcnn::activation_backward_gpu(cudaStream_t, uint32_t, tcnn::Activation, const T *,
  1552. const T *, T *) [with T=tcnn::network_precision_t]"
  1553. (319): here
  1554. instantiation of "void tcnn::activation_backward_gpu(cudaStream_t, tcnn::Activation, const tcnn::GPUMatri
  1555. xDynamic<T> &, tcnn::GPUMatrixDynamic<T> &) [with T=tcnn::network_precision_t]"
  1556. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_mlp.cu(353): here
  1557. instantiation of "void tcnn::CutlassMLP<T>::backward(cudaStream_t, const tcnn::Context &, const tcnn::GPU
  1558. MatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> *
  1559. , __nv_bool, __nv_bool) [with T=tcnn::network_precision_t]"
  1560. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_mlp.cu(452): here
  1561.  
  1562. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  1563. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  1564. detected during:
  1565. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/common_device.h(148): error : more than one conversio
  1566. n function from "tcnn::network_precision_t" to a built-in type applies: [C:\ngp\instant-ngp\build\dependencies\tiny-cud
  1567. a-nn\src\tiny-cuda-nn.vcxproj]
  1568. function "__half::operator float() const"
  1569. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(204): here
  1570. function "__half::operator short() const"
  1571. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(222): here
  1572. function "__half::operator unsigned short() const"
  1573. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(225): here
  1574. function "__half::operator int() const"
  1575. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(228): here
  1576. function "__half::operator unsigned int() const"
  1577. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(231): here
  1578. function "__half::operator long long() const"
  1579. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(234): here
  1580. function "__half::operator unsigned long long() const"
  1581. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(237): here
  1582. function "__half::operator __nv_bool() const"
  1583. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(241): here
  1584. detected during:
  1585. instantiation of "void tcnn::warp_activation_backward_in<T,fragment_t,forward_fragment_t>(tcnn::Activatio
  1586. n, const fragment_t &, const forward_fragment_t &, fragment_t &) [with T=tcnn::network_precision_t, fragment_t=tcnn::
  1587. VectorFragment<tcnn::vector_t<tcnn::network_precision_t, 8U>>, forward_fragment_t=tcnn::VectorFragment<tcnn::vector_t
  1588. <tcnn::network_precision_t, 8U>>]"
  1589. (256): here
  1590. instantiation of "void tcnn::kernel_activation_backward(uint32_t, tcnn::Activation, const T *, const T *,
  1591. T *) [with T=tcnn::network_precision_t, N=8U]"
  1592. (310): here
  1593. instantiation of "void tcnn::activation_backward_gpu(cudaStream_t, uint32_t, tcnn::Activation, const T *,
  1594. const T *, T *) [with T=tcnn::network_precision_t]"
  1595. (319): here
  1596. instantiation of "void tcnn::activation_backward_gpu(cudaStream_t, tcnn::Activation, const tcnn::GPUMatri
  1597. xDynamic<T> &, tcnn::GPUMatrixDynamic<T> &) [with T=tcnn::network_precision_t]"
  1598. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_mlp.cu(353): here
  1599. instantiation of "void tcnn::CutlassMLP<T>::backward(cudaStream_t, const tcnn::Context &, const tcnn::GPU
  1600. MatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> *
  1601. , __nv_bool, __nv_bool) [with T=tcnn::network_precision_t]"
  1602. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_mlp.cu(452): here
  1603.  
  1604. (714): here
  1605.  
  1606. instantiation of "void tcnn::warp_activation<T,fragment_t>(tcnn::Activation, const fragment_t &, fragment
  1607. _t &) [with T=tcnn::network_precision_t, fragment_t=tcnn::VectorFragment<tcnn::vector_t<tcnn::network_precision_t, 8U
  1608. >>]"
  1609. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/common_device.h(156): error : more than one conversio
  1610. n function from "tcnn::network_precision_t" to a built-in type applies: [C:\ngp\instant-ngp\build\dependencies\tiny-cud
  1611. a-nn\src\tiny-cuda-nn.vcxproj]
  1612. function "__half::operator float() const"
  1613. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(204): here
  1614. function "__half::operator short() const"
  1615. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(222): here
  1616. function "__half::operator unsigned short() const"
  1617. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(225): here
  1618. function "__half::operator int() const"
  1619. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(228): here
  1620. function "__half::operator unsigned int() const"
  1621. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(231): here
  1622. function "__half::operator long long() const"
  1623. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(234): here
  1624. function "__half::operator unsigned long long() const"
  1625. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(237): here
  1626. function "__half::operator __nv_bool() const"
  1627. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(241): here
  1628. detected during:
  1629. instantiation of "void tcnn::warp_activation_backward_in<T,fragment_t,forward_fragment_t>(tcnn::Activatio
  1630. n, const fragment_t &, const forward_fragment_t &, fragment_t &) [with T=tcnn::network_precision_t, fragment_t=tcnn::
  1631. VectorFragment<tcnn::vector_t<tcnn::network_precision_t, 8U>>, forward_fragment_t=tcnn::VectorFragment<tcnn::vector_t
  1632. <tcnn::network_precision_t, 8U>>]"
  1633. (256): here
  1634. instantiation of "void tcnn::kernel_activation_backward(uint32_t, tcnn::Activation, const T *, const T *,
  1635. T *) [with T=tcnn::network_precision_t, N=8U]"
  1636. (310): here
  1637. instantiation of "void tcnn::activation_backward_gpu(cudaStream_t, uint32_t, tcnn::Activation, const T *,
  1638. const T *, T *) [with T=tcnn::network_precision_t]"
  1639. (319): here
  1640. instantiation of "void tcnn::activation_backward_gpu(cudaStream_t, tcnn::Activation, const tcnn::GPUMatri
  1641. xDynamic<T> &, tcnn::GPUMatrixDynamic<T> &) [with T=tcnn::network_precision_t]"
  1642. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_mlp.cu(353): here
  1643. instantiation of "void tcnn::CutlassMLP<T>::backward(cudaStream_t, const tcnn::Context &, const tcnn::GPU
  1644. MatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> *
  1645. , __nv_bool, __nv_bool) [with T=tcnn::network_precision_t]"
  1646. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_mlp.cu(452): here
  1647.  
  1648. (244): here
  1649. instantiation of "void tcnn::kernel_activation(uint32_t, tcnn::Activation, const T *, T *) [with T=tcnn::
  1650. network_precision_t, N=8U]"
  1651. (286): here
  1652. instantiation of "void tcnn::activation_gpu(cudaStream_t, uint32_t, tcnn::Activation, const T *, T *) [wi
  1653. th T=tcnn::network_precision_t]"
  1654. (295): here
  1655. instantiation of "void tcnn::activation_gpu(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrixDynamic<
  1656. T> &, tcnn::GPUMatrixDynamic<T> &) [with T=tcnn::network_precision_t]"
  1657. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(66): error : name followed by "::" must be a class
  1658. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  1659. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_resnet.cu(202): here
  1660. instantiation of "std::unique_ptr<tcnn::Context, std::default_delete<tcnn::Context>> tcnn::CutlassResNet<
  1661. T, input_activation>::forward(cudaStream_t, const tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> *, __nv_bool
  1662. , __nv_bool) [with T=tcnn::network_precision_t, input_activation=tcnn::Activation::None]"
  1663. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_resnet.cu(391): here
  1664.  
  1665. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/common_device.h(187): error : more than one conversio
  1666. n function from "tcnn::network_precision_t" to a built-in type applies: [C:\ngp\instant-ngp\build\dependencies\tiny-cud
  1667. a-nn\src\tiny-cuda-nn.vcxproj]
  1668. function "__half::operator float() const"
  1669. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(204): here
  1670. function "__half::operator short() const"
  1671. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/common_device.h(156): error : more than one conversio
  1672. n function from "tcnn::network_precision_t" to a built-in type applies: [C:\ngp\instant-ngp\build\dependencies\tiny-cud
  1673. a-nn\src\tiny-cuda-nn.vcxproj]
  1674. function "__half::operator float() const"
  1675. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(204): here
  1676. function "__half::operator short() const"
  1677. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(222): here
  1678. function "__half::operator unsigned short() const"
  1679. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(225): here
  1680. function "__half::operator int() const"
  1681. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(228): here
  1682. function "__half::operator unsigned int() const"
  1683. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(231): here
  1684. function "__half::operator long long() const"
  1685. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(234): here
  1686. function "__half::operator unsigned long long() const"
  1687. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(237): here
  1688. function "__half::operator __nv_bool() const"
  1689. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(241): here
  1690. detected during:
  1691. instantiation of "void tcnn::warp_activation_backward_in<T,fragment_t,forward_fragment_t>(tcnn::Activatio
  1692. n, const fragment_t &, const forward_fragment_t &, fragment_t &) [with T=tcnn::network_precision_t, fragment_t=tcnn::
  1693. VectorFragment<tcnn::vector_t<tcnn::network_precision_t, 8U>>, forward_fragment_t=tcnn::VectorFragment<tcnn::vector_t
  1694. <tcnn::network_precision_t, 8U>>]"
  1695. (256): here
  1696. instantiation of "void tcnn::kernel_activation_backward(uint32_t, tcnn::Activation, const T *, const T *,
  1697. T *) [with T=tcnn::network_precision_t, N=8U]"
  1698. (310): here
  1699. instantiation of "void tcnn::activation_backward_gpu(cudaStream_t, uint32_t, tcnn::Activation, const T *,
  1700. const T *, T *) [with T=tcnn::network_precision_t]"
  1701. (319): here
  1702. instantiation of "void tcnn::activation_backward_gpu(cudaStream_t, tcnn::Activation, const tcnn::GPUMatri
  1703. xDynamic<T> &, tcnn::GPUMatrixDynamic<T> &) [with T=tcnn::network_precision_t]"
  1704. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_mlp.cu(353): here
  1705. instantiation of "void tcnn::CutlassMLP<T>::backward(cudaStream_t, const tcnn::Context &, const tcnn::GPU
  1706. MatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> *
  1707. , __nv_bool, __nv_bool) [with T=tcnn::network_precision_t]"
  1708. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_mlp.cu(452): here
  1709.  
  1710. detected during:
  1711. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(222): here
  1712. instantiation of "void tcnn::threadblock_layer<WIDTH,N_ITERS,OUT_T,BACKWARD>(tcnn::Activation, __half *,
  1713. const __half *, OUT_T *, const OUT_T *) [with WIDTH=128, N_ITERS=8, OUT_T=__half, BACKWARD=false]"
  1714. function "__half::operator unsigned short() const"
  1715. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(225): here
  1716. function "__half::operator int() const"
  1717. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(228): here
  1718. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/common_device.h(163): error : more than one conversio
  1719. n function from "tcnn::network_precision_t" to a built-in type applies: [C:\ngp\instant-ngp\build\dependencies\tiny-cud
  1720. a-nn\src\tiny-cuda-nn.vcxproj]
  1721. function "__half::operator float() const"
  1722. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(204): here
  1723. function "__half::operator short() const"
  1724. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(222): here
  1725. function "__half::operator unsigned short() const"
  1726. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(225): here
  1727. function "__half::operator int() const"
  1728. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(228): here
  1729. function "__half::operator unsigned int() const"
  1730. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(231): here
  1731. function "__half::operator long long() const"
  1732. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(234): here
  1733. function "__half::operator unsigned long long() const"
  1734. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(237): here
  1735. function "__half::operator __nv_bool() const"
  1736. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(241): here
  1737. detected during:
  1738. instantiation of "void tcnn::warp_activation_backward_in<T,fragment_t,forward_fragment_t>(tcnn::Activatio
  1739. n, const fragment_t &, const forward_fragment_t &, fragment_t &) [with T=tcnn::network_precision_t, fragment_t=tcnn::
  1740. VectorFragment<tcnn::vector_t<tcnn::network_precision_t, 8U>>, forward_fragment_t=tcnn::VectorFragment<tcnn::vector_t
  1741. <tcnn::network_precision_t, 8U>>]"
  1742. (256): here
  1743. instantiation of "void tcnn::kernel_activation_backward(uint32_t, tcnn::Activation, const T *, const T *,
  1744. T *) [with T=tcnn::network_precision_t, N=8U]"
  1745. (310): here
  1746. instantiation of "void tcnn::activation_backward_gpu(cudaStream_t, uint32_t, tcnn::Activation, const T *,
  1747. const T *, T *) [with T=tcnn::network_precision_t]"
  1748. (319): here
  1749. instantiation of "void tcnn::activation_backward_gpu(cudaStream_t, tcnn::Activation, const tcnn::GPUMatri
  1750. xDynamic<T> &, tcnn::GPUMatrixDynamic<T> &) [with T=tcnn::network_precision_t]"
  1751. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_mlp.cu(353): here
  1752. instantiation of "void tcnn::CutlassMLP<T>::backward(cudaStream_t, const tcnn::Context &, const tcnn::GPU
  1753. MatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> *
  1754. , __nv_bool, __nv_bool) [with T=tcnn::network_precision_t]"
  1755. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_mlp.cu(452): here
  1756.  
  1757. function "__half::operator unsigned int() const"
  1758. (525): here
  1759. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(231): here
  1760. instantiation of "void tcnn::kernel_mlp_fused<WIDTH,N_ITERS,OUT_T,ACTIVATION,INFERENCE>(tcnn::Activation,
  1761. const __half *, const __half *, OUT_T *, OUT_T *, uint32_t, uint32_t, uint32_t, uint32_t, <error-type>, <error-type>
  1762. ) [with WIDTH=128, N_ITERS=8, OUT_T=__half, ACTIVATION=tcnn::Activation::None, INFERENCE=true]"
  1763. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/common_device.h(163): error : more than one conversio
  1764. n function from "tcnn::network_precision_t" to a built-in type applies: [C:\ngp\instant-ngp\build\dependencies\tiny-cud
  1765. a-nn\src\tiny-cuda-nn.vcxproj]
  1766. function "__half::operator float() const"
  1767. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(204): here
  1768. function "__half::operator short() const"
  1769. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(222): here
  1770. function "__half::operator unsigned short() const"
  1771. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(225): here
  1772. function "__half::operator int() const"
  1773. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(228): here
  1774. function "__half::operator unsigned int() const"
  1775. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(231): here
  1776. function "__half::operator long long() const"
  1777. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(234): here
  1778. function "__half::operator unsigned long long() const"
  1779. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(237): here
  1780. function "__half::operator __nv_bool() const"
  1781. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(241): here
  1782. detected during:
  1783. instantiation of "void tcnn::warp_activation_backward_in<T,fragment_t,forward_fragment_t>(tcnn::Activatio
  1784. n, const fragment_t &, const forward_fragment_t &, fragment_t &) [with T=tcnn::network_precision_t, fragment_t=tcnn::
  1785. VectorFragment<tcnn::vector_t<tcnn::network_precision_t, 8U>>, forward_fragment_t=tcnn::VectorFragment<tcnn::vector_t
  1786. <tcnn::network_precision_t, 8U>>]"
  1787. (256): here
  1788. instantiation of "void tcnn::kernel_activation_backward(uint32_t, tcnn::Activation, const T *, const T *,
  1789. T *) [with T=tcnn::network_precision_t, N=8U]"
  1790. (310): here
  1791. instantiation of "void tcnn::activation_backward_gpu(cudaStream_t, uint32_t, tcnn::Activation, const T *,
  1792. const T *, T *) [with T=tcnn::network_precision_t]"
  1793. (319): here
  1794. instantiation of "void tcnn::activation_backward_gpu(cudaStream_t, tcnn::Activation, const tcnn::GPUMatri
  1795. xDynamic<T> &, tcnn::GPUMatrixDynamic<T> &) [with T=tcnn::network_precision_t]"
  1796. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_mlp.cu(353): here
  1797. instantiation of "void tcnn::CutlassMLP<T>::backward(cudaStream_t, const tcnn::Context &, const tcnn::GPU
  1798. MatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> *
  1799. , __nv_bool, __nv_bool) [with T=tcnn::network_precision_t]"
  1800. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_mlp.cu(452): here
  1801.  
  1802. function "__half::operator long long() const"
  1803. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(234): here
  1804. (636): here
  1805. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  1806. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  1807. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  1808. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::None, INFERENCE=true]"
  1809. function "__half::operator unsigned long long() const"
  1810. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(237): here
  1811. function "__half::operator __nv_bool() const"
  1812. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(241): here
  1813. detected during:
  1814. instantiation of "void tcnn::warp_activation_backward<T,fragment_t,forward_fragment_t>(tcnn::Activation,
  1815. const fragment_t &, const forward_fragment_t &, fragment_t &) [with T=tcnn::network_precision_t, fragment_t=tcnn::Vec
  1816. torFragment<tcnn::vector_t<tcnn::network_precision_t, 8U>>, forward_fragment_t=tcnn::VectorFragment<tcnn::vector_t<tc
  1817. nn::network_precision_t, 8U>>]"
  1818. (269): here
  1819. instantiation of "void tcnn::kernel_activation_backward_output(uint32_t, tcnn::Activation, const T *, con
  1820. st T *, T *) [with T=tcnn::network_precision_t, N=8U]"
  1821. (334): here
  1822. instantiation of "void tcnn::activation_backward_output_gpu(cudaStream_t, uint32_t, tcnn::Activation, con
  1823. st T *, const T *, T *) [with T=tcnn::network_precision_t]"
  1824. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_resnet.cu(274): here
  1825. instantiation of "void tcnn::CutlassResNet<T, input_activation>::backward(cudaStream_t, const tcnn::Conte
  1826. xt &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, tcnn::
  1827. GPUMatrixDynamic<T> *, __nv_bool, __nv_bool) [with T=tcnn::network_precision_t, input_activation=tcnn::Activation::No
  1828. ne]"
  1829. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_resnet.cu(391): here
  1830.  
  1831. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/common_device.h(187): error : more than one conversio
  1832. n function from "tcnn::network_precision_t" to a built-in type applies: [C:\ngp\instant-ngp\build\dependencies\tiny-cud
  1833. a-nn\src\tiny-cuda-nn.vcxproj]
  1834. function "__half::operator float() const"
  1835. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(204): here
  1836. function "__half::operator short() const"
  1837. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(222): here
  1838. function "__half::operator unsigned short() const"
  1839. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(225): here
  1840. function "__half::operator int() const"
  1841. (760): here
  1842. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(228): here
  1843. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  1844. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  1845. 8 errors detected in the compilation of "C:/ngp/instant-ngp/dependencies/tiny-cuda-nn/src/encoding.cu".
  1846. (714): here
  1847. function "__half::operator unsigned int() const"
  1848. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(231): here
  1849.  
  1850. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(66): error : type name is not allowed [C:\ngp\insta
  1851. nt-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  1852. detected during:
  1853. instantiation of "void tcnn::threadblock_layer<WIDTH,N_ITERS,OUT_T,BACKWARD>(tcnn::Activation, __half *,
  1854. const __half *, OUT_T *, const OUT_T *) [with WIDTH=128, N_ITERS=8, OUT_T=__half, BACKWARD=false]"
  1855. (525): here
  1856. instantiation of "void tcnn::kernel_mlp_fused<WIDTH,N_ITERS,OUT_T,ACTIVATION,INFERENCE>(tcnn::Activation,
  1857. const __half *, const __half *, OUT_T *, OUT_T *, uint32_t, uint32_t, uint32_t, uint32_t, <error-type>, <error-type>
  1858. ) [with WIDTH=128, N_ITERS=8, OUT_T=__half, ACTIVATION=tcnn::Activation::None, INFERENCE=true]"
  1859. function "__half::operator long long() const"
  1860. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(234): here
  1861. function "__half::operator unsigned long long() const"
  1862. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(237): here
  1863. function "__half::operator __nv_bool() const"
  1864. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(241): here
  1865. detected during:
  1866. instantiation of "void tcnn::warp_activation_backward<T,fragment_t,forward_fragment_t>(tcnn::Activation,
  1867. const fragment_t &, const forward_fragment_t &, fragment_t &) [with T=tcnn::network_precision_t, fragment_t=tcnn::Vec
  1868. torFragment<tcnn::vector_t<tcnn::network_precision_t, 8U>>, forward_fragment_t=tcnn::VectorFragment<tcnn::vector_t<tc
  1869. nn::network_precision_t, 8U>>]"
  1870. (269): here
  1871. instantiation of "void tcnn::kernel_activation_backward_output(uint32_t, tcnn::Activation, const T *, con
  1872. st T *, T *) [with T=tcnn::network_precision_t, N=8U]"
  1873. (334): here
  1874. (636): here
  1875. instantiation of "void tcnn::activation_backward_output_gpu(cudaStream_t, uint32_t, tcnn::Activation, con
  1876. st T *, const T *, T *) [with T=tcnn::network_precision_t]"
  1877. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_resnet.cu(274): here
  1878. instantiation of "void tcnn::CutlassResNet<T, input_activation>::backward(cudaStream_t, const tcnn::Conte
  1879. xt &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, tcnn::
  1880. GPUMatrixDynamic<T> *, __nv_bool, __nv_bool) [with T=tcnn::network_precision_t, input_activation=tcnn::Activation::No
  1881. ne]"
  1882. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_resnet.cu(391): here
  1883.  
  1884. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/common_device.h(193): error : more than one conversio
  1885. n function from "tcnn::network_precision_t" to a built-in type applies: [C:\ngp\instant-ngp\build\dependencies\tiny-cud
  1886. a-nn\src\tiny-cuda-nn.vcxproj]
  1887. function "__half::operator float() const"
  1888. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(204): here
  1889. function "__half::operator short() const"
  1890. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(222): here
  1891. function "__half::operator unsigned short() const"
  1892. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(225): here
  1893. function "__half::operator int() const"
  1894. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(228): here
  1895. function "__half::operator unsigned int() const"
  1896. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(231): here
  1897. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  1898. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  1899. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  1900. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::None, INFERENCE=true]"
  1901. function "__half::operator long long() const"
  1902. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(234): here
  1903. function "__half::operator unsigned long long() const"
  1904. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(237): here
  1905. function "__half::operator __nv_bool() const"
  1906. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(241): here
  1907. detected during:
  1908. instantiation of "void tcnn::warp_activation_backward<T,fragment_t,forward_fragment_t>(tcnn::Activation,
  1909. const fragment_t &, const forward_fragment_t &, fragment_t &) [with T=tcnn::network_precision_t, fragment_t=tcnn::Vec
  1910. torFragment<tcnn::vector_t<tcnn::network_precision_t, 8U>>, forward_fragment_t=tcnn::VectorFragment<tcnn::vector_t<tc
  1911. nn::network_precision_t, 8U>>]"
  1912. (269): here
  1913. instantiation of "void tcnn::kernel_activation_backward_output(uint32_t, tcnn::Activation, const T *, con
  1914. st T *, T *) [with T=tcnn::network_precision_t, N=8U]"
  1915. (334): here
  1916. instantiation of "void tcnn::activation_backward_output_gpu(cudaStream_t, uint32_t, tcnn::Activation, con
  1917. st T *, const T *, T *) [with T=tcnn::network_precision_t]"
  1918. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_resnet.cu(274): here
  1919. instantiation of "void tcnn::CutlassResNet<T, input_activation>::backward(cudaStream_t, const tcnn::Conte
  1920. xt &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, tcnn::
  1921. GPUMatrixDynamic<T> *, __nv_bool, __nv_bool) [with T=tcnn::network_precision_t, input_activation=tcnn::Activation::No
  1922. ne]"
  1923. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_resnet.cu(391): here
  1924.  
  1925. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/common_device.h(193): error : more than one conversio
  1926. n function from "tcnn::network_precision_t" to a built-in type applies: [C:\ngp\instant-ngp\build\dependencies\tiny-cud
  1927. a-nn\src\tiny-cuda-nn.vcxproj]
  1928. function "__half::operator float() const"
  1929. (760): here
  1930. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(204): here
  1931. function "__half::operator short() const"
  1932. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(222): here
  1933. function "__half::operator unsigned short() const"
  1934. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  1935. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  1936. (714): here
  1937.  
  1938. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(66): error : name followed by "::" must be a class
  1939. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  1940. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(225): here
  1941. detected during:
  1942. instantiation of "void tcnn::threadblock_layer<WIDTH,N_ITERS,OUT_T,BACKWARD>(tcnn::Activation, __half *,
  1943. const __half *, OUT_T *, const OUT_T *) [with WIDTH=128, N_ITERS=8, OUT_T=__half, BACKWARD=false]"
  1944. (525): here
  1945. instantiation of "void tcnn::kernel_mlp_fused<WIDTH,N_ITERS,OUT_T,ACTIVATION,INFERENCE>(tcnn::Activation,
  1946. const __half *, const __half *, OUT_T *, OUT_T *, uint32_t, uint32_t, uint32_t, uint32_t, <error-type>, <error-type>
  1947. ) [with WIDTH=128, N_ITERS=8, OUT_T=__half, ACTIVATION=tcnn::Activation::None, INFERENCE=true]"
  1948. encoding.cu
  1949. (636): here
  1950. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  1951. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  1952. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  1953. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::None, INFERENCE=true]"
  1954. (760): here
  1955. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  1956. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  1957. (714): here
  1958.  
  1959. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(66): error : identifier "act_frag" is undefined [C:
  1960. \ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  1961. detected during:
  1962. instantiation of "void tcnn::threadblock_layer<WIDTH,N_ITERS,OUT_T,BACKWARD>(tcnn::Activation, __half *,
  1963. const __half *, OUT_T *, const OUT_T *) [with WIDTH=128, N_ITERS=8, OUT_T=__half, BACKWARD=false]"
  1964. (525): here
  1965. instantiation of "void tcnn::kernel_mlp_fused<WIDTH,N_ITERS,OUT_T,ACTIVATION,INFERENCE>(tcnn::Activation,
  1966. const __half *, const __half *, OUT_T *, OUT_T *, uint32_t, uint32_t, uint32_t, uint32_t, <error-type>, <error-type>
  1967. ) [with WIDTH=128, N_ITERS=8, OUT_T=__half, ACTIVATION=tcnn::Activation::None, INFERENCE=true]"
  1968. (636): here
  1969. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  1970. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  1971. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  1972. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::None, INFERENCE=true]"
  1973. (760): here
  1974. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  1975. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  1976. (714): here
  1977.  
  1978. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(67): error : name followed by "::" must be a class
  1979. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  1980. detected during:
  1981. instantiation of "void tcnn::threadblock_layer<WIDTH,N_ITERS,OUT_T,BACKWARD>(tcnn::Activation, __half *,
  1982. const __half *, OUT_T *, const OUT_T *) [with WIDTH=128, N_ITERS=8, OUT_T=__half, BACKWARD=false]"
  1983. (525): here
  1984. instantiation of "void tcnn::kernel_mlp_fused<WIDTH,N_ITERS,OUT_T,ACTIVATION,INFERENCE>(tcnn::Activation,
  1985. const __half *, const __half *, OUT_T *, OUT_T *, uint32_t, uint32_t, uint32_t, uint32_t, <error-type>, <error-type>
  1986. ) [with WIDTH=128, N_ITERS=8, OUT_T=__half, ACTIVATION=tcnn::Activation::None, INFERENCE=true]"
  1987. (636): here
  1988. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  1989. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  1990. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  1991. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::None, INFERENCE=true]"
  1992. (760): here
  1993. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  1994. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  1995. (714): here
  1996.  
  1997. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(67): error : type name is not allowed [C:\ngp\insta
  1998. nt-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  1999. detected during:
  2000. instantiation of "void tcnn::threadblock_layer<WIDTH,N_ITERS,OUT_T,BACKWARD>(tcnn::Activation, __half *,
  2001. const __half *, OUT_T *, const OUT_T *) [with WIDTH=128, N_ITERS=8, OUT_T=__half, BACKWARD=false]"
  2002. function "__half::operator int() const"
  2003. (525): here
  2004. instantiation of "void tcnn::kernel_mlp_fused<WIDTH,N_ITERS,OUT_T,ACTIVATION,INFERENCE>(tcnn::Activation,
  2005. const __half *, const __half *, OUT_T *, OUT_T *, uint32_t, uint32_t, uint32_t, uint32_t, <error-type>, <error-type>
  2006. ) [with WIDTH=128, N_ITERS=8, OUT_T=__half, ACTIVATION=tcnn::Activation::None, INFERENCE=true]"
  2007. (636): here
  2008. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  2009. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  2010. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  2011. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::None, INFERENCE=true]"
  2012. (760): here
  2013. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  2014. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  2015. (714): here
  2016.  
  2017. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(67): error : type name is not allowed [C:\ngp\insta
  2018. nt-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  2019. detected during:
  2020. instantiation of "void tcnn::threadblock_layer<WIDTH,N_ITERS,OUT_T,BACKWARD>(tcnn::Activation, __half *,
  2021. const __half *, OUT_T *, const OUT_T *) [with WIDTH=128, N_ITERS=8, OUT_T=__half, BACKWARD=false]"
  2022. (525): here
  2023. instantiation of "void tcnn::kernel_mlp_fused<WIDTH,N_ITERS,OUT_T,ACTIVATION,INFERENCE>(tcnn::Activation,
  2024. const __half *, const __half *, OUT_T *, OUT_T *, uint32_t, uint32_t, uint32_t, uint32_t, <error-type>, <error-type>
  2025. ) [with WIDTH=128, N_ITERS=8, OUT_T=__half, ACTIVATION=tcnn::Activation::None, INFERENCE=true]"
  2026. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(228): here
  2027. function "__half::operator unsigned int() const"
  2028. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(231): here
  2029. function "__half::operator long long() const"
  2030. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(234): here
  2031. function "__half::operator unsigned long long() const"
  2032. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(237): here
  2033. function "__half::operator __nv_bool() const"
  2034. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(241): here
  2035. detected during:
  2036. instantiation of "void tcnn::warp_activation_backward<T,fragment_t,forward_fragment_t>(tcnn::Activation,
  2037. const fragment_t &, const forward_fragment_t &, fragment_t &) [with T=tcnn::network_precision_t, fragment_t=tcnn::Vec
  2038. torFragment<tcnn::vector_t<tcnn::network_precision_t, 8U>>, forward_fragment_t=tcnn::VectorFragment<tcnn::vector_t<tc
  2039. nn::network_precision_t, 8U>>]"
  2040. (269): here
  2041. instantiation of "void tcnn::kernel_activation_backward_output(uint32_t, tcnn::Activation, const T *, con
  2042. st T *, T *) [with T=tcnn::network_precision_t, N=8U]"
  2043. (334): here
  2044. instantiation of "void tcnn::activation_backward_output_gpu(cudaStream_t, uint32_t, tcnn::Activation, con
  2045. st T *, const T *, T *) [with T=tcnn::network_precision_t]"
  2046. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_resnet.cu(274): here
  2047. instantiation of "void tcnn::CutlassResNet<T, input_activation>::backward(cudaStream_t, const tcnn::Conte
  2048. xt &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, tcnn::
  2049. GPUMatrixDynamic<T> *, __nv_bool, __nv_bool) [with T=tcnn::network_precision_t, input_activation=tcnn::Activation::No
  2050. ne]"
  2051. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_resnet.cu(391): here
  2052.  
  2053. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/common_device.h(204): error : more than one conversio
  2054. n function from "tcnn::network_precision_t" to a built-in type applies: [C:\ngp\instant-ngp\build\dependencies\tiny-cud
  2055. a-nn\src\tiny-cuda-nn.vcxproj]
  2056. function "__half::operator float() const"
  2057. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(204): here
  2058. function "__half::operator short() const"
  2059. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(222): here
  2060. function "__half::operator unsigned short() const"
  2061. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(225): here
  2062. function "__half::operator int() const"
  2063. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(228): here
  2064. function "__half::operator unsigned int() const"
  2065. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(231): here
  2066. function "__half::operator long long() const"
  2067. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(234): here
  2068. function "__half::operator unsigned long long() const"
  2069. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(237): here
  2070. function "__half::operator __nv_bool() const"
  2071. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(241): here
  2072. detected during:
  2073. instantiation of "void tcnn::warp_activation_backward<T,fragment_t,forward_fragment_t>(tcnn::Activation,
  2074. const fragment_t &, const forward_fragment_t &, fragment_t &) [with T=tcnn::network_precision_t, fragment_t=tcnn::Vec
  2075. torFragment<tcnn::vector_t<tcnn::network_precision_t, 8U>>, forward_fragment_t=tcnn::VectorFragment<tcnn::vector_t<tc
  2076. nn::network_precision_t, 8U>>]"
  2077. (269): here
  2078. instantiation of "void tcnn::kernel_activation_backward_output(uint32_t, tcnn::Activation, const T *, con
  2079. st T *, T *) [with T=tcnn::network_precision_t, N=8U]"
  2080. (334): here
  2081. instantiation of "void tcnn::activation_backward_output_gpu(cudaStream_t, uint32_t, tcnn::Activation, con
  2082. st T *, const T *, T *) [with T=tcnn::network_precision_t]"
  2083. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_resnet.cu(274): here
  2084. instantiation of "void tcnn::CutlassResNet<T, input_activation>::backward(cudaStream_t, const tcnn::Conte
  2085. xt &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, tcnn::
  2086. GPUMatrixDynamic<T> *, __nv_bool, __nv_bool) [with T=tcnn::network_precision_t, input_activation=tcnn::Activation::No
  2087. ne]"
  2088. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_resnet.cu(391): here
  2089.  
  2090. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/common_device.h(204): error : more than one conversio
  2091. n function from "tcnn::network_precision_t" to a built-in type applies: [C:\ngp\instant-ngp\build\dependencies\tiny-cud
  2092. a-nn\src\tiny-cuda-nn.vcxproj]
  2093. function "__half::operator float() const"
  2094. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(204): here
  2095. function "__half::operator short() const"
  2096. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(222): here
  2097. function "__half::operator unsigned short() const"
  2098. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(225): here
  2099. function "__half::operator int() const"
  2100. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(228): here
  2101. function "__half::operator unsigned int() const"
  2102. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(231): here
  2103. function "__half::operator long long() const"
  2104. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(234): here
  2105. function "__half::operator unsigned long long() const"
  2106. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(237): here
  2107. function "__half::operator __nv_bool() const"
  2108. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(241): here
  2109. detected during:
  2110. instantiation of "void tcnn::warp_activation_backward<T,fragment_t,forward_fragment_t>(tcnn::Activation,
  2111. const fragment_t &, const forward_fragment_t &, fragment_t &) [with T=tcnn::network_precision_t, fragment_t=tcnn::Vec
  2112. torFragment<tcnn::vector_t<tcnn::network_precision_t, 8U>>, forward_fragment_t=tcnn::VectorFragment<tcnn::vector_t<tc
  2113. nn::network_precision_t, 8U>>]"
  2114. (269): here
  2115. instantiation of "void tcnn::kernel_activation_backward_output(uint32_t, tcnn::Activation, const T *, con
  2116. st T *, T *) [with T=tcnn::network_precision_t, N=8U]"
  2117. (334): here
  2118. instantiation of "void tcnn::activation_backward_output_gpu(cudaStream_t, uint32_t, tcnn::Activation, con
  2119. st T *, const T *, T *) [with T=tcnn::network_precision_t]"
  2120. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_resnet.cu(274): here
  2121. instantiation of "void tcnn::CutlassResNet<T, input_activation>::backward(cudaStream_t, const tcnn::Conte
  2122. xt &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, tcnn::
  2123. GPUMatrixDynamic<T> *, __nv_bool, __nv_bool) [with T=tcnn::network_precision_t, input_activation=tcnn::Activation::No
  2124. ne]"
  2125. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_resnet.cu(391): here
  2126.  
  2127. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/common_device.h(211): error : more than one conversio
  2128. n function from "tcnn::network_precision_t" to a built-in type applies: [C:\ngp\instant-ngp\build\dependencies\tiny-cud
  2129. a-nn\src\tiny-cuda-nn.vcxproj]
  2130. function "__half::operator float() const"
  2131. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(204): here
  2132. function "__half::operator short() const"
  2133. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(222): here
  2134. function "__half::operator unsigned short() const"
  2135. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(225): here
  2136. function "__half::operator int() const"
  2137. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(228): here
  2138. function "__half::operator unsigned int() const"
  2139. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(231): here
  2140. function "__half::operator long long() const"
  2141. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(234): here
  2142. function "__half::operator unsigned long long() const"
  2143. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(237): here
  2144. function "__half::operator __nv_bool() const"
  2145. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(241): here
  2146. detected during:
  2147. instantiation of "void tcnn::warp_activation_backward<T,fragment_t,forward_fragment_t>(tcnn::Activation,
  2148. const fragment_t &, const forward_fragment_t &, fragment_t &) [with T=tcnn::network_precision_t, fragment_t=tcnn::Vec
  2149. torFragment<tcnn::vector_t<tcnn::network_precision_t, 8U>>, forward_fragment_t=tcnn::VectorFragment<tcnn::vector_t<tc
  2150. nn::network_precision_t, 8U>>]"
  2151. (269): here
  2152. instantiation of "void tcnn::kernel_activation_backward_output(uint32_t, tcnn::Activation, const T *, con
  2153. st T *, T *) [with T=tcnn::network_precision_t, N=8U]"
  2154. (334): here
  2155. instantiation of "void tcnn::activation_backward_output_gpu(cudaStream_t, uint32_t, tcnn::Activation, con
  2156. st T *, const T *, T *) [with T=tcnn::network_precision_t]"
  2157. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_resnet.cu(274): here
  2158. instantiation of "void tcnn::CutlassResNet<T, input_activation>::backward(cudaStream_t, const tcnn::Conte
  2159. xt &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, tcnn::
  2160. GPUMatrixDynamic<T> *, __nv_bool, __nv_bool) [with T=tcnn::network_precision_t, input_activation=tcnn::Activation::No
  2161. ne]"
  2162. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_resnet.cu(391): here
  2163.  
  2164. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/common_device.h(211): error : more than one conversio
  2165. n function from "tcnn::network_precision_t" to a built-in type applies: [C:\ngp\instant-ngp\build\dependencies\tiny-cud
  2166. a-nn\src\tiny-cuda-nn.vcxproj]
  2167. function "__half::operator float() const"
  2168. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(204): here
  2169. function "__half::operator short() const"
  2170. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(222): here
  2171. function "__half::operator unsigned short() const"
  2172. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(225): here
  2173. function "__half::operator int() const"
  2174. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(228): here
  2175. function "__half::operator unsigned int() const"
  2176. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(231): here
  2177. function "__half::operator long long() const"
  2178. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(234): here
  2179. function "__half::operator unsigned long long() const"
  2180. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(237): here
  2181. function "__half::operator __nv_bool() const"
  2182. (636): here
  2183. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(241): here
  2184. detected during:
  2185. instantiation of "void tcnn::warp_activation_backward<T,fragment_t,forward_fragment_t>(tcnn::Activation,
  2186. const fragment_t &, const forward_fragment_t &, fragment_t &) [with T=tcnn::network_precision_t, fragment_t=tcnn::Vec
  2187. torFragment<tcnn::vector_t<tcnn::network_precision_t, 8U>>, forward_fragment_t=tcnn::VectorFragment<tcnn::vector_t<tc
  2188. nn::network_precision_t, 8U>>]"
  2189. (269): here
  2190. instantiation of "void tcnn::kernel_activation_backward_output(uint32_t, tcnn::Activation, const T *, con
  2191. st T *, T *) [with T=tcnn::network_precision_t, N=8U]"
  2192. (334): here
  2193. instantiation of "void tcnn::activation_backward_output_gpu(cudaStream_t, uint32_t, tcnn::Activation, con
  2194. st T *, const T *, T *) [with T=tcnn::network_precision_t]"
  2195. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_resnet.cu(274): here
  2196. instantiation of "void tcnn::CutlassResNet<T, input_activation>::backward(cudaStream_t, const tcnn::Conte
  2197. xt &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, tcnn::
  2198. GPUMatrixDynamic<T> *, __nv_bool, __nv_bool) [with T=tcnn::network_precision_t, input_activation=tcnn::Activation::No
  2199. ne]"
  2200. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_resnet.cu(391): here
  2201.  
  2202. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/common_device.h(217): error : more than one conversio
  2203. n function from "tcnn::network_precision_t" to a built-in type applies: [C:\ngp\instant-ngp\build\dependencies\tiny-cud
  2204. a-nn\src\tiny-cuda-nn.vcxproj]
  2205. function "__half::operator float() const"
  2206. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(204): here
  2207. function "__half::operator short() const"
  2208. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(222): here
  2209. function "__half::operator unsigned short() const"
  2210. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(225): here
  2211. function "__half::operator int() const"
  2212. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(228): here
  2213. function "__half::operator unsigned int() const"
  2214. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(231): here
  2215. function "__half::operator long long() const"
  2216. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(234): here
  2217. function "__half::operator unsigned long long() const"
  2218. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(237): here
  2219. function "__half::operator __nv_bool() const"
  2220. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(241): here
  2221. detected during:
  2222. instantiation of "void tcnn::warp_activation_backward<T,fragment_t,forward_fragment_t>(tcnn::Activation,
  2223. const fragment_t &, const forward_fragment_t &, fragment_t &) [with T=tcnn::network_precision_t, fragment_t=tcnn::Vec
  2224. torFragment<tcnn::vector_t<tcnn::network_precision_t, 8U>>, forward_fragment_t=tcnn::VectorFragment<tcnn::vector_t<tc
  2225. nn::network_precision_t, 8U>>]"
  2226. (269): here
  2227. instantiation of "void tcnn::kernel_activation_backward_output(uint32_t, tcnn::Activation, const T *, con
  2228. st T *, T *) [with T=tcnn::network_precision_t, N=8U]"
  2229. (334): here
  2230. instantiation of "void tcnn::activation_backward_output_gpu(cudaStream_t, uint32_t, tcnn::Activation, con
  2231. st T *, const T *, T *) [with T=tcnn::network_precision_t]"
  2232. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_resnet.cu(274): here
  2233. instantiation of "void tcnn::CutlassResNet<T, input_activation>::backward(cudaStream_t, const tcnn::Conte
  2234. xt &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, tcnn::
  2235. GPUMatrixDynamic<T> *, __nv_bool, __nv_bool) [with T=tcnn::network_precision_t, input_activation=tcnn::Activation::No
  2236. ne]"
  2237. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_resnet.cu(391): here
  2238.  
  2239. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/common_device.h(217): error : more than one conversio
  2240. n function from "tcnn::network_precision_t" to a built-in type applies: [C:\ngp\instant-ngp\build\dependencies\tiny-cud
  2241. a-nn\src\tiny-cuda-nn.vcxproj]
  2242. function "__half::operator float() const"
  2243. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(204): here
  2244. function "__half::operator short() const"
  2245. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(222): here
  2246. function "__half::operator unsigned short() const"
  2247. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(225): here
  2248. function "__half::operator int() const"
  2249. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(228): here
  2250. function "__half::operator unsigned int() const"
  2251. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(231): here
  2252. function "__half::operator long long() const"
  2253. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(234): here
  2254. function "__half::operator unsigned long long() const"
  2255. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(237): here
  2256. function "__half::operator __nv_bool() const"
  2257. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(241): here
  2258. detected during:
  2259. instantiation of "void tcnn::warp_activation_backward<T,fragment_t,forward_fragment_t>(tcnn::Activation,
  2260. const fragment_t &, const forward_fragment_t &, fragment_t &) [with T=tcnn::network_precision_t, fragment_t=tcnn::Vec
  2261. torFragment<tcnn::vector_t<tcnn::network_precision_t, 8U>>, forward_fragment_t=tcnn::VectorFragment<tcnn::vector_t<tc
  2262. nn::network_precision_t, 8U>>]"
  2263. (269): here
  2264. instantiation of "void tcnn::kernel_activation_backward_output(uint32_t, tcnn::Activation, const T *, con
  2265. st T *, T *) [with T=tcnn::network_precision_t, N=8U]"
  2266. (334): here
  2267. instantiation of "void tcnn::activation_backward_output_gpu(cudaStream_t, uint32_t, tcnn::Activation, con
  2268. st T *, const T *, T *) [with T=tcnn::network_precision_t]"
  2269. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_resnet.cu(274): here
  2270. instantiation of "void tcnn::CutlassResNet<T, input_activation>::backward(cudaStream_t, const tcnn::Conte
  2271. xt &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, tcnn::
  2272. GPUMatrixDynamic<T> *, __nv_bool, __nv_bool) [with T=tcnn::network_precision_t, input_activation=tcnn::Activation::No
  2273. ne]"
  2274. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_resnet.cu(391): here
  2275.  
  2276. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  2277. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  2278. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  2279. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::None, INFERENCE=true]"
  2280. (760): here
  2281. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  2282. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  2283. (714): here
  2284.  
  2285. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(67): error : identifier "weights_frag" is undefined
  2286. [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  2287. detected during:
  2288. instantiation of "void tcnn::threadblock_layer<WIDTH,N_ITERS,OUT_T,BACKWARD>(tcnn::Activation, __half *,
  2289. const __half *, OUT_T *, const OUT_T *) [with WIDTH=128, N_ITERS=8, OUT_T=__half, BACKWARD=false]"
  2290. (525): here
  2291. instantiation of "void tcnn::kernel_mlp_fused<WIDTH,N_ITERS,OUT_T,ACTIVATION,INFERENCE>(tcnn::Activation,
  2292. const __half *, const __half *, OUT_T *, OUT_T *, uint32_t, uint32_t, uint32_t, uint32_t, <error-type>, <error-type>
  2293. ) [with WIDTH=128, N_ITERS=8, OUT_T=__half, ACTIVATION=tcnn::Activation::None, INFERENCE=true]"
  2294. (636): here
  2295. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  2296. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  2297. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  2298. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::None, INFERENCE=true]"
  2299. (760): here
  2300. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  2301. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  2302. (714): here
  2303. C:\Program Files (x86)\Microsoft Visual Studio\2019\Community\MSBuild\Microsoft\VC\v160\BuildCustomizations\CUDA 11.6.t
  2304. argets(790,9): error MSB3721: The command ""C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\bin\nvcc.exe" -gen
  2305. code=arch=compute_52,code=\"sm_52,compute_52\" --use-local-env -ccbin "C:\Program Files (x86)\Microsoft Visual Studio\2
  2306. 019\Community\VC\Tools\MSVC\14.29.30133\bin\HostX64\x64" -x cu -I"C:\ngp\instant-ngp\dependencies" -I"C:\ngp\instant-
  2307. ngp\dependencies\eigen" -I"C:\ngp\instant-ngp\dependencies\filesystem" -I"C:\ngp\instant-ngp\dependencies\glfw\include"
  2308. -I"C:\ngp\instant-ngp\dependencies\imgui\gl3w" -I"C:\ngp\instant-ngp\dependencies\nanovdb" -I"C:\ProgramData\NVIDIA Co
  2309. rporation\OptiX SDK 7.4.0\include" -I"C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include" -I"C:\ngp\instant-ngp\depen
  2310. dencies\tiny-cuda-nn\dependencies" -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include" --keep-dir
  2311. x64\RelWithDebInfo -maxrregcount=0 --machine 64 --compile -cudart static --extended-lambda --expt-relaxed-constexpr -
  2312. std=c++14 -Xcompiler="/EHsc -Zi -Ob1 -bigobj" -D_WINDOWS -DNDEBUG -DNGP_GUI -DNGP_OPTIX -DTCNN_MIN_GPU_ARCH=75 -DTCNN
  2313. _SHAMPOO -D"CMAKE_INTDIR=\"RelWithDebInfo\"" -D_MBCS -D"CMAKE_INTDIR=\"RelWithDebInfo\"" -Xcompiler "/EHsc /W1 /nologo
  2314. /O2 /FdC:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\RelWithDebInfo\tiny-cuda-nn.pdb /FS /Zi /MD /GR" -o tiny
  2315. -cuda-nn.dir\RelWithDebInfo\encoding.obj "C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\encoding.cu"" exited with co
  2316. de 1. [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  2317.  
  2318. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(68): error : name followed by "::" must be a class
  2319. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  2320. detected during:
  2321. instantiation of "void tcnn::threadblock_layer<WIDTH,N_ITERS,OUT_T,BACKWARD>(tcnn::Activation, __half *,
  2322. const __half *, OUT_T *, const OUT_T *) [with WIDTH=128, N_ITERS=8, OUT_T=__half, BACKWARD=false]"
  2323. (525): here
  2324. instantiation of "void tcnn::kernel_mlp_fused<WIDTH,N_ITERS,OUT_T,ACTIVATION,INFERENCE>(tcnn::Activation,
  2325. const __half *, const __half *, OUT_T *, OUT_T *, uint32_t, uint32_t, uint32_t, uint32_t, <error-type>, <error-type>
  2326. ) [with WIDTH=128, N_ITERS=8, OUT_T=__half, ACTIVATION=tcnn::Activation::None, INFERENCE=true]"
  2327. (636): here
  2328. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  2329. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  2330. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  2331. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::None, INFERENCE=true]"
  2332. (760): here
  2333. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  2334. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  2335. (714): here
  2336.  
  2337. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(68): error : type name is not allowed [C:\ngp\insta
  2338. nt-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  2339. detected during:
  2340. instantiation of "void tcnn::threadblock_layer<WIDTH,N_ITERS,OUT_T,BACKWARD>(tcnn::Activation, __half *,
  2341. const __half *, OUT_T *, const OUT_T *) [with WIDTH=128, N_ITERS=8, OUT_T=__half, BACKWARD=false]"
  2342. (525): here
  2343. instantiation of "void tcnn::kernel_mlp_fused<WIDTH,N_ITERS,OUT_T,ACTIVATION,INFERENCE>(tcnn::Activation,
  2344. const __half *, const __half *, OUT_T *, OUT_T *, uint32_t, uint32_t, uint32_t, uint32_t, <error-type>, <error-type>
  2345. ) [with WIDTH=128, N_ITERS=8, OUT_T=__half, ACTIVATION=tcnn::Activation::None, INFERENCE=true]"
  2346. (636): here
  2347. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  2348. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  2349. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  2350. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::None, INFERENCE=true]"
  2351. (760): here
  2352. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  2353. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  2354. (714): here
  2355.  
  2356. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(68): error : identifier "result_frag" is undefined
  2357. [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  2358. detected during:
  2359. instantiation of "void tcnn::threadblock_layer<WIDTH,N_ITERS,OUT_T,BACKWARD>(tcnn::Activation, __half *,
  2360. const __half *, OUT_T *, const OUT_T *) [with WIDTH=128, N_ITERS=8, OUT_T=__half, BACKWARD=false]"
  2361. (525): here
  2362. instantiation of "void tcnn::kernel_mlp_fused<WIDTH,N_ITERS,OUT_T,ACTIVATION,INFERENCE>(tcnn::Activation,
  2363. const __half *, const __half *, OUT_T *, OUT_T *, uint32_t, uint32_t, uint32_t, uint32_t, <error-type>, <error-type>
  2364. ) [with WIDTH=128, N_ITERS=8, OUT_T=__half, ACTIVATION=tcnn::Activation::None, INFERENCE=true]"
  2365. (636): here
  2366. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  2367. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  2368. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  2369. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::None, INFERENCE=true]"
  2370. (760): here
  2371. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  2372. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  2373. (714): here
  2374.  
  2375. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(87): error : name followed by "::" must be a class
  2376. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  2377. detected during:
  2378. instantiation of "void tcnn::threadblock_layer<WIDTH,N_ITERS,OUT_T,BACKWARD>(tcnn::Activation, __half *,
  2379. const __half *, OUT_T *, const OUT_T *) [with WIDTH=128, N_ITERS=8, OUT_T=__half, BACKWARD=false]"
  2380. (525): here
  2381. instantiation of "void tcnn::kernel_mlp_fused<WIDTH,N_ITERS,OUT_T,ACTIVATION,INFERENCE>(tcnn::Activation,
  2382. const __half *, const __half *, OUT_T *, OUT_T *, uint32_t, uint32_t, uint32_t, uint32_t, <error-type>, <error-type>
  2383. ) [with WIDTH=128, N_ITERS=8, OUT_T=__half, ACTIVATION=tcnn::Activation::None, INFERENCE=true]"
  2384. (636): here
  2385. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  2386. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  2387. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  2388. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::None, INFERENCE=true]"
  2389. (760): here
  2390. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  2391. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  2392. (714): here
  2393.  
  2394. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(89): error : name followed by "::" must be a class
  2395. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  2396. detected during:
  2397. instantiation of "void tcnn::threadblock_layer<WIDTH,N_ITERS,OUT_T,BACKWARD>(tcnn::Activation, __half *,
  2398. const __half *, OUT_T *, const OUT_T *) [with WIDTH=128, N_ITERS=8, OUT_T=__half, BACKWARD=false]"
  2399. (525): here
  2400. instantiation of "void tcnn::kernel_mlp_fused<WIDTH,N_ITERS,OUT_T,ACTIVATION,INFERENCE>(tcnn::Activation,
  2401. const __half *, const __half *, OUT_T *, OUT_T *, uint32_t, uint32_t, uint32_t, uint32_t, <error-type>, <error-type>
  2402. ) [with WIDTH=128, N_ITERS=8, OUT_T=__half, ACTIVATION=tcnn::Activation::None, INFERENCE=true]"
  2403. (636): here
  2404. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  2405. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  2406. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  2407. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::None, INFERENCE=true]"
  2408. (760): here
  2409. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  2410. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  2411. (714): here
  2412.  
  2413. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(95): error : name followed by "::" must be a class
  2414. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  2415. detected during:
  2416. instantiation of "void tcnn::threadblock_layer<WIDTH,N_ITERS,OUT_T,BACKWARD>(tcnn::Activation, __half *,
  2417. const __half *, OUT_T *, const OUT_T *) [with WIDTH=128, N_ITERS=8, OUT_T=__half, BACKWARD=false]"
  2418. (525): here
  2419. instantiation of "void tcnn::kernel_mlp_fused<WIDTH,N_ITERS,OUT_T,ACTIVATION,INFERENCE>(tcnn::Activation,
  2420. const __half *, const __half *, OUT_T *, OUT_T *, uint32_t, uint32_t, uint32_t, uint32_t, <error-type>, <error-type>
  2421. ) [with WIDTH=128, N_ITERS=8, OUT_T=__half, ACTIVATION=tcnn::Activation::None, INFERENCE=true]"
  2422. (636): here
  2423. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  2424. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  2425. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  2426. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::None, INFERENCE=true]"
  2427. (760): here
  2428. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  2429. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  2430. (714): here
  2431.  
  2432. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(100): error : name followed by "::" must be a class
  2433. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  2434. detected during:
  2435. instantiation of "void tcnn::threadblock_layer<WIDTH,N_ITERS,OUT_T,BACKWARD>(tcnn::Activation, __half *,
  2436. const __half *, OUT_T *, const OUT_T *) [with WIDTH=128, N_ITERS=8, OUT_T=__half, BACKWARD=false]"
  2437. (525): here
  2438. instantiation of "void tcnn::kernel_mlp_fused<WIDTH,N_ITERS,OUT_T,ACTIVATION,INFERENCE>(tcnn::Activation,
  2439. const __half *, const __half *, OUT_T *, OUT_T *, uint32_t, uint32_t, uint32_t, uint32_t, <error-type>, <error-type>
  2440. ) [with WIDTH=128, N_ITERS=8, OUT_T=__half, ACTIVATION=tcnn::Activation::None, INFERENCE=true]"
  2441. (636): here
  2442. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  2443. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  2444. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  2445. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::None, INFERENCE=true]"
  2446. (760): here
  2447. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  2448. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  2449. (714): here
  2450.  
  2451. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(101): error : name followed by "::" must be a class
  2452. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  2453. detected during:
  2454. instantiation of "void tcnn::threadblock_layer<WIDTH,N_ITERS,OUT_T,BACKWARD>(tcnn::Activation, __half *,
  2455. const __half *, OUT_T *, const OUT_T *) [with WIDTH=128, N_ITERS=8, OUT_T=__half, BACKWARD=false]"
  2456. (525): here
  2457. instantiation of "void tcnn::kernel_mlp_fused<WIDTH,N_ITERS,OUT_T,ACTIVATION,INFERENCE>(tcnn::Activation,
  2458. const __half *, const __half *, OUT_T *, OUT_T *, uint32_t, uint32_t, uint32_t, uint32_t, <error-type>, <error-type>
  2459. ) [with WIDTH=128, N_ITERS=8, OUT_T=__half, ACTIVATION=tcnn::Activation::None, INFERENCE=true]"
  2460. (636): here
  2461. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  2462. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  2463. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  2464. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::None, INFERENCE=true]"
  2465. (760): here
  2466. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  2467. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  2468. (714): here
  2469.  
  2470. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(107): error : name followed by "::" must be a class
  2471. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  2472. detected during:
  2473. instantiation of "void tcnn::threadblock_layer<WIDTH,N_ITERS,OUT_T,BACKWARD>(tcnn::Activation, __half *,
  2474. const __half *, OUT_T *, const OUT_T *) [with WIDTH=128, N_ITERS=8, OUT_T=__half, BACKWARD=false]"
  2475. (525): here
  2476. instantiation of "void tcnn::kernel_mlp_fused<WIDTH,N_ITERS,OUT_T,ACTIVATION,INFERENCE>(tcnn::Activation,
  2477. const __half *, const __half *, OUT_T *, OUT_T *, uint32_t, uint32_t, uint32_t, uint32_t, <error-type>, <error-type>
  2478. ) [with WIDTH=128, N_ITERS=8, OUT_T=__half, ACTIVATION=tcnn::Activation::None, INFERENCE=true]"
  2479. (636): here
  2480. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  2481. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  2482. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  2483. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::None, INFERENCE=true]"
  2484. (760): here
  2485. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  2486. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  2487. (714): here
  2488.  
  2489. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(118): error : name followed by "::" must be a class
  2490. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  2491. detected during:
  2492. instantiation of "void tcnn::threadblock_layer<WIDTH,N_ITERS,OUT_T,BACKWARD>(tcnn::Activation, __half *,
  2493. const __half *, OUT_T *, const OUT_T *) [with WIDTH=128, N_ITERS=8, OUT_T=__half, BACKWARD=false]"
  2494. (525): here
  2495. instantiation of "void tcnn::kernel_mlp_fused<WIDTH,N_ITERS,OUT_T,ACTIVATION,INFERENCE>(tcnn::Activation,
  2496. const __half *, const __half *, OUT_T *, OUT_T *, uint32_t, uint32_t, uint32_t, uint32_t, <error-type>, <error-type>
  2497. ) [with WIDTH=128, N_ITERS=8, OUT_T=__half, ACTIVATION=tcnn::Activation::None, INFERENCE=true]"
  2498. (636): here
  2499. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  2500. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  2501. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  2502. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::None, INFERENCE=true]"
  2503. (760): here
  2504. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  2505. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  2506. (714): here
  2507.  
  2508. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(118): error : name followed by "::" must be a class
  2509. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  2510. detected during:
  2511. instantiation of "void tcnn::threadblock_layer<WIDTH,N_ITERS,OUT_T,BACKWARD>(tcnn::Activation, __half *,
  2512. const __half *, OUT_T *, const OUT_T *) [with WIDTH=128, N_ITERS=8, OUT_T=__half, BACKWARD=false]"
  2513. (525): here
  2514. instantiation of "void tcnn::kernel_mlp_fused<WIDTH,N_ITERS,OUT_T,ACTIVATION,INFERENCE>(tcnn::Activation,
  2515. const __half *, const __half *, OUT_T *, OUT_T *, uint32_t, uint32_t, uint32_t, uint32_t, <error-type>, <error-type>
  2516. ) [with WIDTH=128, N_ITERS=8, OUT_T=__half, ACTIVATION=tcnn::Activation::None, INFERENCE=true]"
  2517. (636): here
  2518. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  2519. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  2520. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  2521. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::None, INFERENCE=true]"
  2522. (760): here
  2523. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  2524. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  2525. (714): here
  2526.  
  2527. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(544): error : name followed by "::" must be a class
  2528. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  2529. detected during:
  2530. instantiation of "void tcnn::kernel_mlp_fused<WIDTH,N_ITERS,OUT_T,ACTIVATION,INFERENCE>(tcnn::Activation,
  2531. const __half *, const __half *, OUT_T *, OUT_T *, uint32_t, uint32_t, uint32_t, uint32_t, <error-type>, <error-type>
  2532. ) [with WIDTH=128, N_ITERS=8, OUT_T=__half, ACTIVATION=tcnn::Activation::None, INFERENCE=true]"
  2533. (636): here
  2534. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  2535. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  2536. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  2537. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::None, INFERENCE=true]"
  2538. (760): here
  2539. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  2540. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  2541. (714): here
  2542.  
  2543. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(634): error : name followed by "::" must be a class
  2544. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  2545. detected during:
  2546. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  2547. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  2548. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  2549. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::Exponential, INFERENCE=true]"
  2550. (761): here
  2551. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  2552. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  2553. (714): here
  2554.  
  2555. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(634): error : name followed by "::" must be a class
  2556. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  2557. detected during:
  2558. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  2559. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  2560. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  2561. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::Exponential, INFERENCE=true]"
  2562. (761): here
  2563. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  2564. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  2565. (714): here
  2566.  
  2567. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(635): error : name followed by "::" must be a class
  2568. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  2569. detected during:
  2570. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  2571. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  2572. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  2573. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::Exponential, INFERENCE=true]"
  2574. (761): here
  2575. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  2576. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  2577. (714): here
  2578.  
  2579. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(635): error : name followed by "::" must be a class
  2580. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  2581. detected during:
  2582. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  2583. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  2584. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  2585. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::Exponential, INFERENCE=true]"
  2586. (761): here
  2587. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  2588. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  2589. (714): here
  2590.  
  2591. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(515): error : name followed by "::" must be a class
  2592. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  2593. detected during:
  2594. instantiation of "void tcnn::kernel_mlp_fused<WIDTH,N_ITERS,OUT_T,ACTIVATION,INFERENCE>(tcnn::Activation,
  2595. const __half *, const __half *, OUT_T *, OUT_T *, uint32_t, uint32_t, uint32_t, uint32_t, <error-type>, <error-type>
  2596. ) [with WIDTH=128, N_ITERS=8, OUT_T=__half, ACTIVATION=tcnn::Activation::Exponential, INFERENCE=true]"
  2597. (636): here
  2598. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  2599. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  2600. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  2601. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::Exponential, INFERENCE=true]"
  2602. (761): here
  2603. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  2604. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  2605. (714): here
  2606.  
  2607. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(516): error : name followed by "::" must be a class
  2608. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  2609. detected during:
  2610. instantiation of "void tcnn::kernel_mlp_fused<WIDTH,N_ITERS,OUT_T,ACTIVATION,INFERENCE>(tcnn::Activation,
  2611. const __half *, const __half *, OUT_T *, OUT_T *, uint32_t, uint32_t, uint32_t, uint32_t, <error-type>, <error-type>
  2612. ) [with WIDTH=128, N_ITERS=8, OUT_T=__half, ACTIVATION=tcnn::Activation::Exponential, INFERENCE=true]"
  2613. (636): here
  2614. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  2615. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  2616. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  2617. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::Exponential, INFERENCE=true]"
  2618. (761): here
  2619. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  2620. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  2621. (714): here
  2622.  
  2623. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(517): error : name followed by "::" must be a class
  2624. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  2625. detected during:
  2626. instantiation of "void tcnn::kernel_mlp_fused<WIDTH,N_ITERS,OUT_T,ACTIVATION,INFERENCE>(tcnn::Activation,
  2627. const __half *, const __half *, OUT_T *, OUT_T *, uint32_t, uint32_t, uint32_t, uint32_t, <error-type>, <error-type>
  2628. ) [with WIDTH=128, N_ITERS=8, OUT_T=__half, ACTIVATION=tcnn::Activation::Exponential, INFERENCE=true]"
  2629. (636): here
  2630. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  2631. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  2632. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  2633. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::Exponential, INFERENCE=true]"
  2634. (761): here
  2635. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  2636. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  2637. (714): here
  2638.  
  2639. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(519): error : name followed by "::" must be a class
  2640. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  2641. detected during:
  2642. instantiation of "void tcnn::kernel_mlp_fused<WIDTH,N_ITERS,OUT_T,ACTIVATION,INFERENCE>(tcnn::Activation,
  2643. const __half *, const __half *, OUT_T *, OUT_T *, uint32_t, uint32_t, uint32_t, uint32_t, <error-type>, <error-type>
  2644. ) [with WIDTH=128, N_ITERS=8, OUT_T=__half, ACTIVATION=tcnn::Activation::Exponential, INFERENCE=true]"
  2645. (636): here
  2646. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  2647. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  2648. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  2649. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::Exponential, INFERENCE=true]"
  2650. (761): here
  2651. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  2652. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  2653. (714): here
  2654.  
  2655. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(544): error : name followed by "::" must be a class
  2656. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  2657. detected during:
  2658. instantiation of "void tcnn::kernel_mlp_fused<WIDTH,N_ITERS,OUT_T,ACTIVATION,INFERENCE>(tcnn::Activation,
  2659. const __half *, const __half *, OUT_T *, OUT_T *, uint32_t, uint32_t, uint32_t, uint32_t, <error-type>, <error-type>
  2660. ) [with WIDTH=128, N_ITERS=8, OUT_T=__half, ACTIVATION=tcnn::Activation::Exponential, INFERENCE=true]"
  2661. (636): here
  2662. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  2663. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  2664. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  2665. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::Exponential, INFERENCE=true]"
  2666. (761): here
  2667. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  2668. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  2669. (714): here
  2670.  
  2671. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(634): error : name followed by "::" must be a class
  2672. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  2673. detected during:
  2674. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  2675. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  2676. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  2677. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::Sigmoid, INFERENCE=true]"
  2678. (762): here
  2679. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  2680. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  2681. (714): here
  2682.  
  2683. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(634): error : name followed by "::" must be a class
  2684. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  2685. detected during:
  2686. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  2687. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  2688. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  2689. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::Sigmoid, INFERENCE=true]"
  2690. (762): here
  2691. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  2692. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  2693. (714): here
  2694.  
  2695. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(635): error : name followed by "::" must be a class
  2696. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  2697. detected during:
  2698. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  2699. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  2700. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  2701. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::Sigmoid, INFERENCE=true]"
  2702. (762): here
  2703. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  2704. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  2705. (714): here
  2706.  
  2707. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(635): error : name followed by "::" must be a class
  2708. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  2709. detected during:
  2710. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  2711. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  2712. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  2713. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::Sigmoid, INFERENCE=true]"
  2714. (762): here
  2715. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  2716. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  2717. (714): here
  2718.  
  2719. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(515): error : name followed by "::" must be a class
  2720. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  2721. detected during:
  2722. instantiation of "void tcnn::kernel_mlp_fused<WIDTH,N_ITERS,OUT_T,ACTIVATION,INFERENCE>(tcnn::Activation,
  2723. const __half *, const __half *, OUT_T *, OUT_T *, uint32_t, uint32_t, uint32_t, uint32_t, <error-type>, <error-type>
  2724. ) [with WIDTH=128, N_ITERS=8, OUT_T=__half, ACTIVATION=tcnn::Activation::Sigmoid, INFERENCE=true]"
  2725. (636): here
  2726. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  2727. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  2728. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  2729. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::Sigmoid, INFERENCE=true]"
  2730. (762): here
  2731. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  2732. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  2733. (714): here
  2734.  
  2735. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(516): error : name followed by "::" must be a class
  2736. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  2737. detected during:
  2738. instantiation of "void tcnn::kernel_mlp_fused<WIDTH,N_ITERS,OUT_T,ACTIVATION,INFERENCE>(tcnn::Activation,
  2739. const __half *, const __half *, OUT_T *, OUT_T *, uint32_t, uint32_t, uint32_t, uint32_t, <error-type>, <error-type>
  2740. ) [with WIDTH=128, N_ITERS=8, OUT_T=__half, ACTIVATION=tcnn::Activation::Sigmoid, INFERENCE=true]"
  2741. (636): here
  2742. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  2743. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  2744. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  2745. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::Sigmoid, INFERENCE=true]"
  2746. (762): here
  2747. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  2748. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  2749. (714): here
  2750.  
  2751. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(517): error : name followed by "::" must be a class
  2752. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  2753. detected during:
  2754. instantiation of "void tcnn::kernel_mlp_fused<WIDTH,N_ITERS,OUT_T,ACTIVATION,INFERENCE>(tcnn::Activation,
  2755. const __half *, const __half *, OUT_T *, OUT_T *, uint32_t, uint32_t, uint32_t, uint32_t, <error-type>, <error-type>
  2756. ) [with WIDTH=128, N_ITERS=8, OUT_T=__half, ACTIVATION=tcnn::Activation::Sigmoid, INFERENCE=true]"
  2757. (636): here
  2758. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  2759. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  2760. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  2761. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::Sigmoid, INFERENCE=true]"
  2762. (762): here
  2763. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  2764. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  2765. (714): here
  2766.  
  2767. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(519): error : name followed by "::" must be a class
  2768. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  2769. detected during:
  2770. instantiation of "void tcnn::kernel_mlp_fused<WIDTH,N_ITERS,OUT_T,ACTIVATION,INFERENCE>(tcnn::Activation,
  2771. const __half *, const __half *, OUT_T *, OUT_T *, uint32_t, uint32_t, uint32_t, uint32_t, <error-type>, <error-type>
  2772. ) [with WIDTH=128, N_ITERS=8, OUT_T=__half, ACTIVATION=tcnn::Activation::Sigmoid, INFERENCE=true]"
  2773. (636): here
  2774. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  2775. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  2776. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  2777. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::Sigmoid, INFERENCE=true]"
  2778. (762): here
  2779. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  2780. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  2781. (714): here
  2782.  
  2783. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(544): error : name followed by "::" must be a class
  2784. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  2785. detected during:
  2786. instantiation of "void tcnn::kernel_mlp_fused<WIDTH,N_ITERS,OUT_T,ACTIVATION,INFERENCE>(tcnn::Activation,
  2787. const __half *, const __half *, OUT_T *, OUT_T *, uint32_t, uint32_t, uint32_t, uint32_t, <error-type>, <error-type>
  2788. ) [with WIDTH=128, N_ITERS=8, OUT_T=__half, ACTIVATION=tcnn::Activation::Sigmoid, INFERENCE=true]"
  2789. (636): here
  2790. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  2791. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  2792. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  2793. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::Sigmoid, INFERENCE=true]"
  2794. (762): here
  2795. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  2796. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  2797. (714): here
  2798.  
  2799. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(634): error : name followed by "::" must be a class
  2800. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  2801. detected during:
  2802. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  2803. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  2804. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  2805. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::ReLU, INFERENCE=true]"
  2806. (763): here
  2807. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  2808. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  2809. (714): here
  2810.  
  2811. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(634): error : name followed by "::" must be a class
  2812. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  2813. detected during:
  2814. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  2815. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  2816. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  2817. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::ReLU, INFERENCE=true]"
  2818. (763): here
  2819. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  2820. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  2821. (714): here
  2822.  
  2823. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(635): error : name followed by "::" must be a class
  2824. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  2825. detected during:
  2826. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  2827. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  2828. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  2829. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::ReLU, INFERENCE=true]"
  2830. (763): here
  2831. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  2832. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  2833. (714): here
  2834.  
  2835. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(635): error : name followed by "::" must be a class
  2836. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  2837. detected during:
  2838. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  2839. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  2840. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  2841. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::ReLU, INFERENCE=true]"
  2842. (763): here
  2843. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  2844. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  2845. (714): here
  2846.  
  2847. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(515): error : name followed by "::" must be a class
  2848. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  2849. detected during:
  2850. instantiation of "void tcnn::kernel_mlp_fused<WIDTH,N_ITERS,OUT_T,ACTIVATION,INFERENCE>(tcnn::Activation,
  2851. const __half *, const __half *, OUT_T *, OUT_T *, uint32_t, uint32_t, uint32_t, uint32_t, <error-type>, <error-type>
  2852. ) [with WIDTH=128, N_ITERS=8, OUT_T=__half, ACTIVATION=tcnn::Activation::ReLU, INFERENCE=true]"
  2853. (636): here
  2854. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  2855. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  2856. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  2857. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::ReLU, INFERENCE=true]"
  2858. (763): here
  2859. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  2860. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  2861. (714): here
  2862.  
  2863. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(516): error : name followed by "::" must be a class
  2864. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  2865. detected during:
  2866. instantiation of "void tcnn::kernel_mlp_fused<WIDTH,N_ITERS,OUT_T,ACTIVATION,INFERENCE>(tcnn::Activation,
  2867. const __half *, const __half *, OUT_T *, OUT_T *, uint32_t, uint32_t, uint32_t, uint32_t, <error-type>, <error-type>
  2868. ) [with WIDTH=128, N_ITERS=8, OUT_T=__half, ACTIVATION=tcnn::Activation::ReLU, INFERENCE=true]"
  2869. (636): here
  2870. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  2871. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  2872. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  2873. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::ReLU, INFERENCE=true]"
  2874. (763): here
  2875. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  2876. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  2877. (714): here
  2878.  
  2879. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(517): error : name followed by "::" must be a class
  2880. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  2881. detected during:
  2882. instantiation of "void tcnn::kernel_mlp_fused<WIDTH,N_ITERS,OUT_T,ACTIVATION,INFERENCE>(tcnn::Activation,
  2883. const __half *, const __half *, OUT_T *, OUT_T *, uint32_t, uint32_t, uint32_t, uint32_t, <error-type>, <error-type>
  2884. ) [with WIDTH=128, N_ITERS=8, OUT_T=__half, ACTIVATION=tcnn::Activation::ReLU, INFERENCE=true]"
  2885. (636): here
  2886. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  2887. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  2888. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  2889. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::ReLU, INFERENCE=true]"
  2890. (763): here
  2891. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  2892. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  2893. (714): here
  2894.  
  2895. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(519): error : name followed by "::" must be a class
  2896. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  2897. detected during:
  2898. instantiation of "void tcnn::kernel_mlp_fused<WIDTH,N_ITERS,OUT_T,ACTIVATION,INFERENCE>(tcnn::Activation,
  2899. const __half *, const __half *, OUT_T *, OUT_T *, uint32_t, uint32_t, uint32_t, uint32_t, <error-type>, <error-type>
  2900. ) [with WIDTH=128, N_ITERS=8, OUT_T=__half, ACTIVATION=tcnn::Activation::ReLU, INFERENCE=true]"
  2901. (636): here
  2902. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  2903. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  2904. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  2905. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::ReLU, INFERENCE=true]"
  2906. (763): here
  2907. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  2908. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  2909. (714): here
  2910.  
  2911. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(544): error : name followed by "::" must be a class
  2912. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  2913. detected during:
  2914. instantiation of "void tcnn::kernel_mlp_fused<WIDTH,N_ITERS,OUT_T,ACTIVATION,INFERENCE>(tcnn::Activation,
  2915. const __half *, const __half *, OUT_T *, OUT_T *, uint32_t, uint32_t, uint32_t, uint32_t, <error-type>, <error-type>
  2916. ) [with WIDTH=128, N_ITERS=8, OUT_T=__half, ACTIVATION=tcnn::Activation::ReLU, INFERENCE=true]"
  2917. (636): here
  2918. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  2919. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  2920. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  2921. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::ReLU, INFERENCE=true]"
  2922. (763): here
  2923. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  2924. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  2925. (714): here
  2926.  
  2927. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(634): error : name followed by "::" must be a class
  2928. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  2929. detected during:
  2930. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  2931. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  2932. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  2933. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::Squareplus, INFERENCE=true]"
  2934. (764): here
  2935. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  2936. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  2937. (714): here
  2938.  
  2939. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(634): error : name followed by "::" must be a class
  2940. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  2941. detected during:
  2942. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  2943. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  2944. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  2945. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::Squareplus, INFERENCE=true]"
  2946. (764): here
  2947. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  2948. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  2949. (714): here
  2950.  
  2951. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(635): error : name followed by "::" must be a class
  2952. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  2953. detected during:
  2954. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  2955. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  2956. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  2957. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::Squareplus, INFERENCE=true]"
  2958. (764): here
  2959. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  2960. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  2961. (714): here
  2962.  
  2963. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(635): error : name followed by "::" must be a class
  2964. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  2965. detected during:
  2966. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  2967. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  2968. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  2969. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::Squareplus, INFERENCE=true]"
  2970. (764): here
  2971. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  2972. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  2973. (714): here
  2974.  
  2975. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(515): error : name followed by "::" must be a class
  2976. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  2977. detected during:
  2978. instantiation of "void tcnn::kernel_mlp_fused<WIDTH,N_ITERS,OUT_T,ACTIVATION,INFERENCE>(tcnn::Activation,
  2979. const __half *, const __half *, OUT_T *, OUT_T *, uint32_t, uint32_t, uint32_t, uint32_t, <error-type>, <error-type>
  2980. ) [with WIDTH=128, N_ITERS=8, OUT_T=__half, ACTIVATION=tcnn::Activation::Squareplus, INFERENCE=true]"
  2981. (636): here
  2982. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  2983. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  2984. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  2985. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::Squareplus, INFERENCE=true]"
  2986. (764): here
  2987. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  2988. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  2989. (714): here
  2990.  
  2991. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(516): error : name followed by "::" must be a class
  2992. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  2993. detected during:
  2994. instantiation of "void tcnn::kernel_mlp_fused<WIDTH,N_ITERS,OUT_T,ACTIVATION,INFERENCE>(tcnn::Activation,
  2995. const __half *, const __half *, OUT_T *, OUT_T *, uint32_t, uint32_t, uint32_t, uint32_t, <error-type>, <error-type>
  2996. ) [with WIDTH=128, N_ITERS=8, OUT_T=__half, ACTIVATION=tcnn::Activation::Squareplus, INFERENCE=true]"
  2997. (636): here
  2998. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  2999. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  3000. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  3001. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::Squareplus, INFERENCE=true]"
  3002. (764): here
  3003. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  3004. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  3005. (714): here
  3006.  
  3007. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(517): error : name followed by "::" must be a class
  3008. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  3009. detected during:
  3010. instantiation of "void tcnn::kernel_mlp_fused<WIDTH,N_ITERS,OUT_T,ACTIVATION,INFERENCE>(tcnn::Activation,
  3011. const __half *, const __half *, OUT_T *, OUT_T *, uint32_t, uint32_t, uint32_t, uint32_t, <error-type>, <error-type>
  3012. ) [with WIDTH=128, N_ITERS=8, OUT_T=__half, ACTIVATION=tcnn::Activation::Squareplus, INFERENCE=true]"
  3013. (636): here
  3014. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  3015. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  3016. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  3017. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::Squareplus, INFERENCE=true]"
  3018. (764): here
  3019. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  3020. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  3021. (714): here
  3022.  
  3023. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(519): error : name followed by "::" must be a class
  3024. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  3025. detected during:
  3026. instantiation of "void tcnn::kernel_mlp_fused<WIDTH,N_ITERS,OUT_T,ACTIVATION,INFERENCE>(tcnn::Activation,
  3027. const __half *, const __half *, OUT_T *, OUT_T *, uint32_t, uint32_t, uint32_t, uint32_t, <error-type>, <error-type>
  3028. ) [with WIDTH=128, N_ITERS=8, OUT_T=__half, ACTIVATION=tcnn::Activation::Squareplus, INFERENCE=true]"
  3029. (636): here
  3030. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  3031. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  3032. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  3033. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::Squareplus, INFERENCE=true]"
  3034. (764): here
  3035. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  3036. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  3037. (714): here
  3038.  
  3039. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(544): error : name followed by "::" must be a class
  3040. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  3041. detected during:
  3042. instantiation of "void tcnn::kernel_mlp_fused<WIDTH,N_ITERS,OUT_T,ACTIVATION,INFERENCE>(tcnn::Activation,
  3043. const __half *, const __half *, OUT_T *, OUT_T *, uint32_t, uint32_t, uint32_t, uint32_t, <error-type>, <error-type>
  3044. ) [with WIDTH=128, N_ITERS=8, OUT_T=__half, ACTIVATION=tcnn::Activation::Squareplus, INFERENCE=true]"
  3045. (636): here
  3046. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  3047. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  3048. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  3049. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::Squareplus, INFERENCE=true]"
  3050. (764): here
  3051. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  3052. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  3053. (714): here
  3054.  
  3055. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(634): error : name followed by "::" must be a class
  3056. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  3057. detected during:
  3058. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  3059. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  3060. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  3061. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::Softplus, INFERENCE=true]"
  3062. (765): here
  3063. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  3064. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  3065. (714): here
  3066.  
  3067. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(634): error : name followed by "::" must be a class
  3068. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  3069. detected during:
  3070. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  3071. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  3072. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  3073. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::Softplus, INFERENCE=true]"
  3074. (765): here
  3075. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  3076. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  3077. (714): here
  3078.  
  3079. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(635): error : name followed by "::" must be a class
  3080. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  3081. detected during:
  3082. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  3083. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  3084. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  3085. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::Softplus, INFERENCE=true]"
  3086. (765): here
  3087. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  3088. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  3089. (714): here
  3090.  
  3091. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(635): error : name followed by "::" must be a class
  3092. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  3093. detected during:
  3094. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  3095. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  3096. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  3097. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::Softplus, INFERENCE=true]"
  3098. (765): here
  3099. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  3100. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  3101. (714): here
  3102.  
  3103. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(515): error : name followed by "::" must be a class
  3104. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  3105. detected during:
  3106. instantiation of "void tcnn::kernel_mlp_fused<WIDTH,N_ITERS,OUT_T,ACTIVATION,INFERENCE>(tcnn::Activation,
  3107. const __half *, const __half *, OUT_T *, OUT_T *, uint32_t, uint32_t, uint32_t, uint32_t, <error-type>, <error-type>
  3108. ) [with WIDTH=128, N_ITERS=8, OUT_T=__half, ACTIVATION=tcnn::Activation::Softplus, INFERENCE=true]"
  3109. (636): here
  3110. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  3111. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  3112. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  3113. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::Softplus, INFERENCE=true]"
  3114. (765): here
  3115. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  3116. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  3117. (714): here
  3118.  
  3119. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(516): error : name followed by "::" must be a class
  3120. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  3121. detected during:
  3122. instantiation of "void tcnn::kernel_mlp_fused<WIDTH,N_ITERS,OUT_T,ACTIVATION,INFERENCE>(tcnn::Activation,
  3123. const __half *, const __half *, OUT_T *, OUT_T *, uint32_t, uint32_t, uint32_t, uint32_t, <error-type>, <error-type>
  3124. ) [with WIDTH=128, N_ITERS=8, OUT_T=__half, ACTIVATION=tcnn::Activation::Softplus, INFERENCE=true]"
  3125. (636): here
  3126. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  3127. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  3128. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  3129. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::Softplus, INFERENCE=true]"
  3130. (765): here
  3131. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  3132. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  3133. (714): here
  3134.  
  3135. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(517): error : name followed by "::" must be a class
  3136. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  3137. detected during:
  3138. instantiation of "void tcnn::kernel_mlp_fused<WIDTH,N_ITERS,OUT_T,ACTIVATION,INFERENCE>(tcnn::Activation,
  3139. const __half *, const __half *, OUT_T *, OUT_T *, uint32_t, uint32_t, uint32_t, uint32_t, <error-type>, <error-type>
  3140. ) [with WIDTH=128, N_ITERS=8, OUT_T=__half, ACTIVATION=tcnn::Activation::Softplus, INFERENCE=true]"
  3141. (636): here
  3142. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  3143. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  3144. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  3145. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::Softplus, INFERENCE=true]"
  3146. (765): here
  3147. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  3148. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  3149. (714): here
  3150.  
  3151. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(519): error : name followed by "::" must be a class
  3152. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  3153. detected during:
  3154. instantiation of "void tcnn::kernel_mlp_fused<WIDTH,N_ITERS,OUT_T,ACTIVATION,INFERENCE>(tcnn::Activation,
  3155. const __half *, const __half *, OUT_T *, OUT_T *, uint32_t, uint32_t, uint32_t, uint32_t, <error-type>, <error-type>
  3156. ) [with WIDTH=128, N_ITERS=8, OUT_T=__half, ACTIVATION=tcnn::Activation::Softplus, INFERENCE=true]"
  3157. (636): here
  3158. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  3159. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  3160. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  3161. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::Softplus, INFERENCE=true]"
  3162. (765): here
  3163. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  3164. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  3165. (714): here
  3166.  
  3167. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(544): error : name followed by "::" must be a class
  3168. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  3169. detected during:
  3170. instantiation of "void tcnn::kernel_mlp_fused<WIDTH,N_ITERS,OUT_T,ACTIVATION,INFERENCE>(tcnn::Activation,
  3171. const __half *, const __half *, OUT_T *, OUT_T *, uint32_t, uint32_t, uint32_t, uint32_t, <error-type>, <error-type>
  3172. ) [with WIDTH=128, N_ITERS=8, OUT_T=__half, ACTIVATION=tcnn::Activation::Softplus, INFERENCE=true]"
  3173. (636): here
  3174. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  3175. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  3176. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  3177. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::Softplus, INFERENCE=true]"
  3178. (765): here
  3179. instantiation of "void tcnn::FullyFusedMLP<T, WIDTH>::inference_mixed_precision(cudaStream_t, const tcnn:
  3180. :GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> &, __nv_bool) [with T=tcnn::network_precision_t, WIDTH=128]"
  3181. (714): here
  3182.  
  3183. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(634): error : name followed by "::" must be a class
  3184. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  3185. detected during:
  3186. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  3187. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  3188. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  3189. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::None, INFERENCE=false]"
  3190. (799): here
  3191. instantiation of "std::unique_ptr<tcnn::Context, std::default_delete<tcnn::Context>> tcnn::FullyFusedMLP<
  3192. T, WIDTH>::forward(cudaStream_t, const tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> *, __nv_bool, __nv_bool
  3193. ) [with T=tcnn::network_precision_t, WIDTH=128]"
  3194. (998): here
  3195.  
  3196. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(634): error : name followed by "::" must be a class
  3197. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  3198. detected during:
  3199. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  3200. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  3201. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  3202. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::None, INFERENCE=false]"
  3203. (799): here
  3204. instantiation of "std::unique_ptr<tcnn::Context, std::default_delete<tcnn::Context>> tcnn::FullyFusedMLP<
  3205. T, WIDTH>::forward(cudaStream_t, const tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> *, __nv_bool, __nv_bool
  3206. ) [with T=tcnn::network_precision_t, WIDTH=128]"
  3207. (998): here
  3208.  
  3209. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(635): error : name followed by "::" must be a class
  3210. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  3211. detected during:
  3212. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  3213. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  3214. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  3215. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::None, INFERENCE=false]"
  3216. (799): here
  3217. instantiation of "std::unique_ptr<tcnn::Context, std::default_delete<tcnn::Context>> tcnn::FullyFusedMLP<
  3218. T, WIDTH>::forward(cudaStream_t, const tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> *, __nv_bool, __nv_bool
  3219. ) [with T=tcnn::network_precision_t, WIDTH=128]"
  3220. (998): here
  3221.  
  3222. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(635): error : name followed by "::" must be a class
  3223. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  3224. detected during:
  3225. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  3226. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  3227. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  3228. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::None, INFERENCE=false]"
  3229. (799): here
  3230. instantiation of "std::unique_ptr<tcnn::Context, std::default_delete<tcnn::Context>> tcnn::FullyFusedMLP<
  3231. T, WIDTH>::forward(cudaStream_t, const tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> *, __nv_bool, __nv_bool
  3232. ) [with T=tcnn::network_precision_t, WIDTH=128]"
  3233. (998): here
  3234.  
  3235. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(515): error : name followed by "::" must be a class
  3236. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  3237. detected during:
  3238. instantiation of "void tcnn::kernel_mlp_fused<WIDTH,N_ITERS,OUT_T,ACTIVATION,INFERENCE>(tcnn::Activation,
  3239. const __half *, const __half *, OUT_T *, OUT_T *, uint32_t, uint32_t, uint32_t, uint32_t, <error-type>, <error-type>
  3240. ) [with WIDTH=128, N_ITERS=8, OUT_T=__half, ACTIVATION=tcnn::Activation::None, INFERENCE=false]"
  3241. (636): here
  3242. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  3243. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  3244. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  3245. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::None, INFERENCE=false]"
  3246. (799): here
  3247. instantiation of "std::unique_ptr<tcnn::Context, std::default_delete<tcnn::Context>> tcnn::FullyFusedMLP<
  3248. T, WIDTH>::forward(cudaStream_t, const tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> *, __nv_bool, __nv_bool
  3249. ) [with T=tcnn::network_precision_t, WIDTH=128]"
  3250. (998): here
  3251.  
  3252. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(516): error : name followed by "::" must be a class
  3253. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  3254. detected during:
  3255. instantiation of "void tcnn::kernel_mlp_fused<WIDTH,N_ITERS,OUT_T,ACTIVATION,INFERENCE>(tcnn::Activation,
  3256. const __half *, const __half *, OUT_T *, OUT_T *, uint32_t, uint32_t, uint32_t, uint32_t, <error-type>, <error-type>
  3257. ) [with WIDTH=128, N_ITERS=8, OUT_T=__half, ACTIVATION=tcnn::Activation::None, INFERENCE=false]"
  3258. (636): here
  3259. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  3260. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  3261. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  3262. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::None, INFERENCE=false]"
  3263. (799): here
  3264. instantiation of "std::unique_ptr<tcnn::Context, std::default_delete<tcnn::Context>> tcnn::FullyFusedMLP<
  3265. T, WIDTH>::forward(cudaStream_t, const tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> *, __nv_bool, __nv_bool
  3266. ) [with T=tcnn::network_precision_t, WIDTH=128]"
  3267. (998): here
  3268.  
  3269. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(517): error : name followed by "::" must be a class
  3270. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  3271. detected during:
  3272. instantiation of "void tcnn::kernel_mlp_fused<WIDTH,N_ITERS,OUT_T,ACTIVATION,INFERENCE>(tcnn::Activation,
  3273. const __half *, const __half *, OUT_T *, OUT_T *, uint32_t, uint32_t, uint32_t, uint32_t, <error-type>, <error-type>
  3274. ) [with WIDTH=128, N_ITERS=8, OUT_T=__half, ACTIVATION=tcnn::Activation::None, INFERENCE=false]"
  3275. (636): here
  3276. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  3277. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  3278. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  3279. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::None, INFERENCE=false]"
  3280. (799): here
  3281. instantiation of "std::unique_ptr<tcnn::Context, std::default_delete<tcnn::Context>> tcnn::FullyFusedMLP<
  3282. T, WIDTH>::forward(cudaStream_t, const tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> *, __nv_bool, __nv_bool
  3283. ) [with T=tcnn::network_precision_t, WIDTH=128]"
  3284. (998): here
  3285.  
  3286. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(519): error : name followed by "::" must be a class
  3287. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  3288. detected during:
  3289. instantiation of "void tcnn::kernel_mlp_fused<WIDTH,N_ITERS,OUT_T,ACTIVATION,INFERENCE>(tcnn::Activation,
  3290. const __half *, const __half *, OUT_T *, OUT_T *, uint32_t, uint32_t, uint32_t, uint32_t, <error-type>, <error-type>
  3291. ) [with WIDTH=128, N_ITERS=8, OUT_T=__half, ACTIVATION=tcnn::Activation::None, INFERENCE=false]"
  3292. (636): here
  3293. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  3294. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  3295. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  3296. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::None, INFERENCE=false]"
  3297. (799): here
  3298. instantiation of "std::unique_ptr<tcnn::Context, std::default_delete<tcnn::Context>> tcnn::FullyFusedMLP<
  3299. T, WIDTH>::forward(cudaStream_t, const tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> *, __nv_bool, __nv_bool
  3300. ) [with T=tcnn::network_precision_t, WIDTH=128]"
  3301. (998): here
  3302.  
  3303. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(544): error : name followed by "::" must be a class
  3304. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  3305. detected during:
  3306. instantiation of "void tcnn::kernel_mlp_fused<WIDTH,N_ITERS,OUT_T,ACTIVATION,INFERENCE>(tcnn::Activation,
  3307. const __half *, const __half *, OUT_T *, OUT_T *, uint32_t, uint32_t, uint32_t, uint32_t, <error-type>, <error-type>
  3308. ) [with WIDTH=128, N_ITERS=8, OUT_T=__half, ACTIVATION=tcnn::Activation::None, INFERENCE=false]"
  3309. (636): here
  3310. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  3311. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  3312. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  3313. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::None, INFERENCE=false]"
  3314. (799): here
  3315. instantiation of "std::unique_ptr<tcnn::Context, std::default_delete<tcnn::Context>> tcnn::FullyFusedMLP<
  3316. T, WIDTH>::forward(cudaStream_t, const tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> *, __nv_bool, __nv_bool
  3317. ) [with T=tcnn::network_precision_t, WIDTH=128]"
  3318. (998): here
  3319.  
  3320. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(634): error : name followed by "::" must be a class
  3321. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  3322. detected during:
  3323. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  3324. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  3325. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  3326. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::Exponential, INFERENCE=false]"
  3327. (800): here
  3328. instantiation of "std::unique_ptr<tcnn::Context, std::default_delete<tcnn::Context>> tcnn::FullyFusedMLP<
  3329. T, WIDTH>::forward(cudaStream_t, const tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> *, __nv_bool, __nv_bool
  3330. ) [with T=tcnn::network_precision_t, WIDTH=128]"
  3331. (998): here
  3332.  
  3333. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(634): error : name followed by "::" must be a class
  3334. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  3335. detected during:
  3336. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  3337. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  3338. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  3339. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::Exponential, INFERENCE=false]"
  3340. (800): here
  3341. instantiation of "std::unique_ptr<tcnn::Context, std::default_delete<tcnn::Context>> tcnn::FullyFusedMLP<
  3342. T, WIDTH>::forward(cudaStream_t, const tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> *, __nv_bool, __nv_bool
  3343. ) [with T=tcnn::network_precision_t, WIDTH=128]"
  3344. (998): here
  3345.  
  3346. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(635): error : name followed by "::" must be a class
  3347. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  3348. detected during:
  3349. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  3350. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  3351. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  3352. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::Exponential, INFERENCE=false]"
  3353. (800): here
  3354. instantiation of "std::unique_ptr<tcnn::Context, std::default_delete<tcnn::Context>> tcnn::FullyFusedMLP<
  3355. T, WIDTH>::forward(cudaStream_t, const tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> *, __nv_bool, __nv_bool
  3356. ) [with T=tcnn::network_precision_t, WIDTH=128]"
  3357. (998): here
  3358.  
  3359. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(635): error : name followed by "::" must be a class
  3360. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  3361. detected during:
  3362. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  3363. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  3364. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  3365. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::Exponential, INFERENCE=false]"
  3366. (800): here
  3367. instantiation of "std::unique_ptr<tcnn::Context, std::default_delete<tcnn::Context>> tcnn::FullyFusedMLP<
  3368. T, WIDTH>::forward(cudaStream_t, const tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> *, __nv_bool, __nv_bool
  3369. ) [with T=tcnn::network_precision_t, WIDTH=128]"
  3370. (998): here
  3371.  
  3372. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(515): error : name followed by "::" must be a class
  3373. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  3374. detected during:
  3375. instantiation of "void tcnn::kernel_mlp_fused<WIDTH,N_ITERS,OUT_T,ACTIVATION,INFERENCE>(tcnn::Activation,
  3376. const __half *, const __half *, OUT_T *, OUT_T *, uint32_t, uint32_t, uint32_t, uint32_t, <error-type>, <error-type>
  3377. ) [with WIDTH=128, N_ITERS=8, OUT_T=__half, ACTIVATION=tcnn::Activation::Exponential, INFERENCE=false]"
  3378. (636): here
  3379. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  3380. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  3381. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  3382. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::Exponential, INFERENCE=false]"
  3383. (800): here
  3384. instantiation of "std::unique_ptr<tcnn::Context, std::default_delete<tcnn::Context>> tcnn::FullyFusedMLP<
  3385. T, WIDTH>::forward(cudaStream_t, const tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> *, __nv_bool, __nv_bool
  3386. ) [with T=tcnn::network_precision_t, WIDTH=128]"
  3387. (998): here
  3388.  
  3389. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(516): error : name followed by "::" must be a class
  3390. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  3391. detected during:
  3392. instantiation of "void tcnn::kernel_mlp_fused<WIDTH,N_ITERS,OUT_T,ACTIVATION,INFERENCE>(tcnn::Activation,
  3393. const __half *, const __half *, OUT_T *, OUT_T *, uint32_t, uint32_t, uint32_t, uint32_t, <error-type>, <error-type>
  3394. ) [with WIDTH=128, N_ITERS=8, OUT_T=__half, ACTIVATION=tcnn::Activation::Exponential, INFERENCE=false]"
  3395. (636): here
  3396. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  3397. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  3398. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  3399. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::Exponential, INFERENCE=false]"
  3400. (800): here
  3401. instantiation of "std::unique_ptr<tcnn::Context, std::default_delete<tcnn::Context>> tcnn::FullyFusedMLP<
  3402. T, WIDTH>::forward(cudaStream_t, const tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> *, __nv_bool, __nv_bool
  3403. ) [with T=tcnn::network_precision_t, WIDTH=128]"
  3404. (998): here
  3405.  
  3406. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(517): error : name followed by "::" must be a class
  3407. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  3408. detected during:
  3409. instantiation of "void tcnn::kernel_mlp_fused<WIDTH,N_ITERS,OUT_T,ACTIVATION,INFERENCE>(tcnn::Activation,
  3410. const __half *, const __half *, OUT_T *, OUT_T *, uint32_t, uint32_t, uint32_t, uint32_t, <error-type>, <error-type>
  3411. ) [with WIDTH=128, N_ITERS=8, OUT_T=__half, ACTIVATION=tcnn::Activation::Exponential, INFERENCE=false]"
  3412. (636): here
  3413. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  3414. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  3415. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  3416. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::Exponential, INFERENCE=false]"
  3417. (800): here
  3418. instantiation of "std::unique_ptr<tcnn::Context, std::default_delete<tcnn::Context>> tcnn::FullyFusedMLP<
  3419. T, WIDTH>::forward(cudaStream_t, const tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> *, __nv_bool, __nv_bool
  3420. ) [with T=tcnn::network_precision_t, WIDTH=128]"
  3421. (998): here
  3422.  
  3423. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/common_device.h(129): error : more than one conversio
  3424. n function from "tcnn::network_precision_t" to a built-in type applies: [C:\ngp\instant-ngp\build\dependencies\tiny-cud
  3425. a-nn\src\tiny-cuda-nn.vcxproj]
  3426. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(519): error : name followed by "::" must be a class
  3427. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  3428. function "__half::operator float() const"
  3429. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(204): here
  3430. function "__half::operator short() const"
  3431. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(222): here
  3432. function "__half::operator unsigned short() const"
  3433. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(225): here
  3434. function "__half::operator int() const"
  3435. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(228): here
  3436. function "__half::operator unsigned int() const"
  3437. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(231): here
  3438. function "__half::operator long long() const"
  3439. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(234): here
  3440. function "__half::operator unsigned long long() const"
  3441. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(237): here
  3442. function "__half::operator __nv_bool() const"
  3443. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(241): here
  3444. detected during:
  3445. instantiation of "void tcnn::warp_activation_backward_in<T,fragment_t,forward_fragment_t>(tcnn::Activatio
  3446. n, const fragment_t &, const forward_fragment_t &, fragment_t &) [with T=tcnn::network_precision_t, fragment_t=tcnn::
  3447. VectorFragment<tcnn::vector_t<tcnn::network_precision_t, 8U>>, forward_fragment_t=tcnn::VectorFragment<tcnn::vector_t
  3448. <tcnn::network_precision_t, 8U>>]"
  3449. (256): here
  3450. instantiation of "void tcnn::kernel_activation_backward(uint32_t, tcnn::Activation, const T *, const T *,
  3451. T *) [with T=tcnn::network_precision_t, N=8U]"
  3452. (310): here
  3453. instantiation of "void tcnn::activation_backward_gpu(cudaStream_t, uint32_t, tcnn::Activation, const T *,
  3454. const T *, T *) [with T=tcnn::network_precision_t]"
  3455. (319): here
  3456. instantiation of "void tcnn::activation_backward_gpu(cudaStream_t, tcnn::Activation, const tcnn::GPUMatri
  3457. xDynamic<T> &, tcnn::GPUMatrixDynamic<T> &) [with T=tcnn::network_precision_t]"
  3458. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_resnet.cu(323): here
  3459. instantiation of "void tcnn::CutlassResNet<T, input_activation>::backward(cudaStream_t, const tcnn::Conte
  3460. xt &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, tcnn::
  3461. GPUMatrixDynamic<T> *, __nv_bool, __nv_bool) [with T=tcnn::network_precision_t, input_activation=tcnn::Activation::No
  3462. ne]"
  3463. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_resnet.cu(391): here
  3464.  
  3465. detected during:
  3466. instantiation of "void tcnn::kernel_mlp_fused<WIDTH,N_ITERS,OUT_T,ACTIVATION,INFERENCE>(tcnn::Activation,
  3467. const __half *, const __half *, OUT_T *, OUT_T *, uint32_t, uint32_t, uint32_t, uint32_t, <error-type>, <error-type>
  3468. ) [with WIDTH=128, N_ITERS=8, OUT_T=__half, ACTIVATION=tcnn::Activation::Exponential, INFERENCE=false]"
  3469. (636): here
  3470. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  3471. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  3472. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  3473. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::Exponential, INFERENCE=false]"
  3474. (800): here
  3475. instantiation of "std::unique_ptr<tcnn::Context, std::default_delete<tcnn::Context>> tcnn::FullyFusedMLP<
  3476. T, WIDTH>::forward(cudaStream_t, const tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> *, __nv_bool, __nv_bool
  3477. ) [with T=tcnn::network_precision_t, WIDTH=128]"
  3478. (998): here
  3479.  
  3480. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(544): error : name followed by "::" must be a class
  3481. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  3482. detected during:
  3483. instantiation of "void tcnn::kernel_mlp_fused<WIDTH,N_ITERS,OUT_T,ACTIVATION,INFERENCE>(tcnn::Activation,
  3484. const __half *, const __half *, OUT_T *, OUT_T *, uint32_t, uint32_t, uint32_t, uint32_t, <error-type>, <error-type>
  3485. ) [with WIDTH=128, N_ITERS=8, OUT_T=__half, ACTIVATION=tcnn::Activation::Exponential, INFERENCE=false]"
  3486. (636): here
  3487. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  3488. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  3489. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  3490. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::Exponential, INFERENCE=false]"
  3491. (800): here
  3492. instantiation of "std::unique_ptr<tcnn::Context, std::default_delete<tcnn::Context>> tcnn::FullyFusedMLP<
  3493. T, WIDTH>::forward(cudaStream_t, const tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> *, __nv_bool, __nv_bool
  3494. ) [with T=tcnn::network_precision_t, WIDTH=128]"
  3495. (998): here
  3496.  
  3497. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(634): error : name followed by "::" must be a class
  3498. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  3499. detected during:
  3500. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  3501. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  3502. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  3503. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::Sigmoid, INFERENCE=false]"
  3504. (801): here
  3505. instantiation of "std::unique_ptr<tcnn::Context, std::default_delete<tcnn::Context>> tcnn::FullyFusedMLP<
  3506. T, WIDTH>::forward(cudaStream_t, const tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> *, __nv_bool, __nv_bool
  3507. ) [with T=tcnn::network_precision_t, WIDTH=128]"
  3508. (998): here
  3509.  
  3510. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(634): error : name followed by "::" must be a class
  3511. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  3512. detected during:
  3513. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  3514. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  3515. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  3516. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::Sigmoid, INFERENCE=false]"
  3517. (801): here
  3518. instantiation of "std::unique_ptr<tcnn::Context, std::default_delete<tcnn::Context>> tcnn::FullyFusedMLP<
  3519. T, WIDTH>::forward(cudaStream_t, const tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> *, __nv_bool, __nv_bool
  3520. ) [with T=tcnn::network_precision_t, WIDTH=128]"
  3521. (998): here
  3522.  
  3523. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu(635): error : name followed by "::" must be a class
  3524. or namespace name [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  3525. detected during:
  3526. instantiation of "std::enable_if_t<std::is_same<__half, T>::value, void> tcnn::mlp_fused_forward<WIDTH,T,
  3527. ACTIVATION,INFERENCE>(cudaStream_t, tcnn::Activation, const tcnn::GPUMatrix<T, tcnn::MatrixLayout::RowMajor> &, const
  3528. tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrix<T, tcnn::MatrixLayout::ColumnMajor> &, tcnn::GPUMatrixDynamic<T> *, uin
  3529. t32_t) [with WIDTH=128, T=tcnn::network_precision_t, ACTIVATION=tcnn::Activation::Sigmoid, INFERENCE=false]"
  3530. (801): here
  3531. instantiation of "std::unique_ptr<tcnn::Context, std::default_delete<tcnn::Context>> tcnn::FullyFusedMLP<
  3532. T, WIDTH>::forward(cudaStream_t, const tcnn::GPUMatrixDynamic<T> &, tcnn::GPUMatrixDynamic<T> *, __nv_bool, __nv_bool
  3533. ) [with T=tcnn::network_precision_t, WIDTH=128]"
  3534. (998): here
  3535.  
  3536. Error limit reached.
  3537. 100 errors detected in the compilation of "C:/ngp/instant-ngp/dependencies/tiny-cuda-nn/src/fully_fused_mlp.cu".
  3538. Compilation terminated.
  3539. fully_fused_mlp.cu
  3540. C:\Program Files (x86)\Microsoft Visual Studio\2019\Community\MSBuild\Microsoft\VC\v160\BuildCustomizations\CUDA 11.6.t
  3541. argets(790,9): error MSB3721: The command ""C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\bin\nvcc.exe" -gen
  3542. code=arch=compute_52,code=\"sm_52,compute_52\" --use-local-env -ccbin "C:\Program Files (x86)\Microsoft Visual Studio\2
  3543. 019\Community\VC\Tools\MSVC\14.29.30133\bin\HostX64\x64" -x cu -I"C:\ngp\instant-ngp\dependencies" -I"C:\ngp\instant-
  3544. ngp\dependencies\eigen" -I"C:\ngp\instant-ngp\dependencies\filesystem" -I"C:\ngp\instant-ngp\dependencies\glfw\include"
  3545. -I"C:\ngp\instant-ngp\dependencies\imgui\gl3w" -I"C:\ngp\instant-ngp\dependencies\nanovdb" -I"C:\ProgramData\NVIDIA Co
  3546. rporation\OptiX SDK 7.4.0\include" -I"C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include" -I"C:\ngp\instant-ngp\depen
  3547. dencies\tiny-cuda-nn\dependencies" -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include" --keep-dir
  3548. x64\RelWithDebInfo -maxrregcount=0 --machine 64 --compile -cudart static --extended-lambda --expt-relaxed-constexpr -
  3549. std=c++14 -Xcompiler="/EHsc -Zi -Ob1 -bigobj" -D_WINDOWS -DNDEBUG -DNGP_GUI -DNGP_OPTIX -DTCNN_MIN_GPU_ARCH=75 -DTCNN
  3550. _SHAMPOO -D"CMAKE_INTDIR=\"RelWithDebInfo\"" -D_MBCS -D"CMAKE_INTDIR=\"RelWithDebInfo\"" -Xcompiler "/EHsc /W1 /nologo
  3551. /O2 /FdC:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\RelWithDebInfo\tiny-cuda-nn.pdb /FS /Zi /MD /GR" -o tiny
  3552. -cuda-nn.dir\RelWithDebInfo\fully_fused_mlp.obj "C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\fully_fused_mlp.cu""
  3553. exited with code 1. [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  3554. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/common_device.h(129): error : more than one conversio
  3555. n function from "tcnn::network_precision_t" to a built-in type applies: [C:\ngp\instant-ngp\build\dependencies\tiny-cud
  3556. a-nn\src\tiny-cuda-nn.vcxproj]
  3557. function "__half::operator float() const"
  3558. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(204): here
  3559. function "__half::operator short() const"
  3560. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(222): here
  3561. function "__half::operator unsigned short() const"
  3562. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(225): here
  3563. function "__half::operator int() const"
  3564. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(228): here
  3565. function "__half::operator unsigned int() const"
  3566. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(231): here
  3567. function "__half::operator long long() const"
  3568. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(234): here
  3569. function "__half::operator unsigned long long() const"
  3570. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(237): here
  3571. function "__half::operator __nv_bool() const"
  3572. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(241): here
  3573. detected during:
  3574. instantiation of "void tcnn::warp_activation_backward_in<T,fragment_t,forward_fragment_t>(tcnn::Activatio
  3575. n, const fragment_t &, const forward_fragment_t &, fragment_t &) [with T=tcnn::network_precision_t, fragment_t=tcnn::
  3576. VectorFragment<tcnn::vector_t<tcnn::network_precision_t, 8U>>, forward_fragment_t=tcnn::VectorFragment<tcnn::vector_t
  3577. <tcnn::network_precision_t, 8U>>]"
  3578. (256): here
  3579. instantiation of "void tcnn::kernel_activation_backward(uint32_t, tcnn::Activation, const T *, const T *,
  3580. T *) [with T=tcnn::network_precision_t, N=8U]"
  3581. (310): here
  3582. instantiation of "void tcnn::activation_backward_gpu(cudaStream_t, uint32_t, tcnn::Activation, const T *,
  3583. const T *, T *) [with T=tcnn::network_precision_t]"
  3584. (319): here
  3585. instantiation of "void tcnn::activation_backward_gpu(cudaStream_t, tcnn::Activation, const tcnn::GPUMatri
  3586. xDynamic<T> &, tcnn::GPUMatrixDynamic<T> &) [with T=tcnn::network_precision_t]"
  3587. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_resnet.cu(323): here
  3588. instantiation of "void tcnn::CutlassResNet<T, input_activation>::backward(cudaStream_t, const tcnn::Conte
  3589. xt &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, tcnn::
  3590. GPUMatrixDynamic<T> *, __nv_bool, __nv_bool) [with T=tcnn::network_precision_t, input_activation=tcnn::Activation::No
  3591. ne]"
  3592. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_resnet.cu(391): here
  3593.  
  3594. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/common_device.h(135): error : more than one conversio
  3595. n function from "tcnn::network_precision_t" to a built-in type applies: [C:\ngp\instant-ngp\build\dependencies\tiny-cud
  3596. a-nn\src\tiny-cuda-nn.vcxproj]
  3597. function "__half::operator float() const"
  3598. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(204): here
  3599. function "__half::operator short() const"
  3600. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(222): here
  3601. function "__half::operator unsigned short() const"
  3602. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(225): here
  3603. function "__half::operator int() const"
  3604. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(228): here
  3605. function "__half::operator unsigned int() const"
  3606. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(231): here
  3607. function "__half::operator long long() const"
  3608. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(234): here
  3609. function "__half::operator unsigned long long() const"
  3610. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(237): here
  3611. function "__half::operator __nv_bool() const"
  3612. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(241): here
  3613. detected during:
  3614. instantiation of "void tcnn::warp_activation_backward_in<T,fragment_t,forward_fragment_t>(tcnn::Activatio
  3615. n, const fragment_t &, const forward_fragment_t &, fragment_t &) [with T=tcnn::network_precision_t, fragment_t=tcnn::
  3616. VectorFragment<tcnn::vector_t<tcnn::network_precision_t, 8U>>, forward_fragment_t=tcnn::VectorFragment<tcnn::vector_t
  3617. <tcnn::network_precision_t, 8U>>]"
  3618. (256): here
  3619. instantiation of "void tcnn::kernel_activation_backward(uint32_t, tcnn::Activation, const T *, const T *,
  3620. T *) [with T=tcnn::network_precision_t, N=8U]"
  3621. (310): here
  3622. instantiation of "void tcnn::activation_backward_gpu(cudaStream_t, uint32_t, tcnn::Activation, const T *,
  3623. const T *, T *) [with T=tcnn::network_precision_t]"
  3624. (319): here
  3625. instantiation of "void tcnn::activation_backward_gpu(cudaStream_t, tcnn::Activation, const tcnn::GPUMatri
  3626. xDynamic<T> &, tcnn::GPUMatrixDynamic<T> &) [with T=tcnn::network_precision_t]"
  3627. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_resnet.cu(323): here
  3628. instantiation of "void tcnn::CutlassResNet<T, input_activation>::backward(cudaStream_t, const tcnn::Conte
  3629. xt &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, tcnn::
  3630. GPUMatrixDynamic<T> *, __nv_bool, __nv_bool) [with T=tcnn::network_precision_t, input_activation=tcnn::Activation::No
  3631. ne]"
  3632. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_resnet.cu(391): here
  3633.  
  3634. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/common_device.h(135): error : more than one conversio
  3635. n function from "tcnn::network_precision_t" to a built-in type applies: [C:\ngp\instant-ngp\build\dependencies\tiny-cud
  3636. a-nn\src\tiny-cuda-nn.vcxproj]
  3637. function "__half::operator float() const"
  3638. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(204): here
  3639. function "__half::operator short() const"
  3640. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(222): here
  3641. function "__half::operator unsigned short() const"
  3642. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(225): here
  3643. function "__half::operator int() const"
  3644. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(228): here
  3645. function "__half::operator unsigned int() const"
  3646. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(231): here
  3647. function "__half::operator long long() const"
  3648. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(234): here
  3649. function "__half::operator unsigned long long() const"
  3650. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(237): here
  3651. function "__half::operator __nv_bool() const"
  3652. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(241): here
  3653. detected during:
  3654. instantiation of "void tcnn::warp_activation_backward_in<T,fragment_t,forward_fragment_t>(tcnn::Activatio
  3655. n, const fragment_t &, const forward_fragment_t &, fragment_t &) [with T=tcnn::network_precision_t, fragment_t=tcnn::
  3656. VectorFragment<tcnn::vector_t<tcnn::network_precision_t, 8U>>, forward_fragment_t=tcnn::VectorFragment<tcnn::vector_t
  3657. <tcnn::network_precision_t, 8U>>]"
  3658. (256): here
  3659. instantiation of "void tcnn::kernel_activation_backward(uint32_t, tcnn::Activation, const T *, const T *,
  3660. T *) [with T=tcnn::network_precision_t, N=8U]"
  3661. (310): here
  3662. instantiation of "void tcnn::activation_backward_gpu(cudaStream_t, uint32_t, tcnn::Activation, const T *,
  3663. const T *, T *) [with T=tcnn::network_precision_t]"
  3664. (319): here
  3665. instantiation of "void tcnn::activation_backward_gpu(cudaStream_t, tcnn::Activation, const tcnn::GPUMatri
  3666. xDynamic<T> &, tcnn::GPUMatrixDynamic<T> &) [with T=tcnn::network_precision_t]"
  3667. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_resnet.cu(323): here
  3668. instantiation of "void tcnn::CutlassResNet<T, input_activation>::backward(cudaStream_t, const tcnn::Conte
  3669. xt &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, tcnn::
  3670. GPUMatrixDynamic<T> *, __nv_bool, __nv_bool) [with T=tcnn::network_precision_t, input_activation=tcnn::Activation::No
  3671. ne]"
  3672. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_resnet.cu(391): here
  3673.  
  3674. 24 errors detected in the compilation of "C:/ngp/instant-ngp/dependencies/tiny-cuda-nn/src/cutlass_mlp.cu".
  3675. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/common_device.h(141): error : more than one conversio
  3676. n function from "tcnn::network_precision_t" to a built-in type applies: [C:\ngp\instant-ngp\build\dependencies\tiny-cud
  3677. a-nn\src\tiny-cuda-nn.vcxproj]
  3678. function "__half::operator float() const"
  3679. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(204): here
  3680. function "__half::operator short() const"
  3681. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(222): here
  3682. function "__half::operator unsigned short() const"
  3683. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(225): here
  3684. function "__half::operator int() const"
  3685. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(228): here
  3686. function "__half::operator unsigned int() const"
  3687. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(231): here
  3688. function "__half::operator long long() const"
  3689. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(234): here
  3690. function "__half::operator unsigned long long() const"
  3691. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(237): here
  3692. function "__half::operator __nv_bool() const"
  3693. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(241): here
  3694. detected during:
  3695. instantiation of "void tcnn::warp_activation_backward_in<T,fragment_t,forward_fragment_t>(tcnn::Activatio
  3696. n, const fragment_t &, const forward_fragment_t &, fragment_t &) [with T=tcnn::network_precision_t, fragment_t=tcnn::
  3697. VectorFragment<tcnn::vector_t<tcnn::network_precision_t, 8U>>, forward_fragment_t=tcnn::VectorFragment<tcnn::vector_t
  3698. <tcnn::network_precision_t, 8U>>]"
  3699. (256): here
  3700. instantiation of "void tcnn::kernel_activation_backward(uint32_t, tcnn::Activation, const T *, const T *,
  3701. T *) [with T=tcnn::network_precision_t, N=8U]"
  3702. (310): here
  3703. instantiation of "void tcnn::activation_backward_gpu(cudaStream_t, uint32_t, tcnn::Activation, const T *,
  3704. const T *, T *) [with T=tcnn::network_precision_t]"
  3705. (319): here
  3706. instantiation of "void tcnn::activation_backward_gpu(cudaStream_t, tcnn::Activation, const tcnn::GPUMatri
  3707. xDynamic<T> &, tcnn::GPUMatrixDynamic<T> &) [with T=tcnn::network_precision_t]"
  3708. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_resnet.cu(323): here
  3709. instantiation of "void tcnn::CutlassResNet<T, input_activation>::backward(cudaStream_t, const tcnn::Conte
  3710. xt &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, tcnn::
  3711. GPUMatrixDynamic<T> *, __nv_bool, __nv_bool) [with T=tcnn::network_precision_t, input_activation=tcnn::Activation::No
  3712. ne]"
  3713. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_resnet.cu(391): here
  3714.  
  3715. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/common_device.h(141): error : more than one conversio
  3716. n function from "tcnn::network_precision_t" to a built-in type applies: [C:\ngp\instant-ngp\build\dependencies\tiny-cud
  3717. a-nn\src\tiny-cuda-nn.vcxproj]
  3718. function "__half::operator float() const"
  3719. cutlass_mlp.cu
  3720. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(204): here
  3721. function "__half::operator short() const"
  3722. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(222): here
  3723. function "__half::operator unsigned short() const"
  3724. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(225): here
  3725. function "__half::operator int() const"
  3726. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(228): here
  3727. function "__half::operator unsigned int() const"
  3728. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(231): here
  3729. function "__half::operator long long() const"
  3730. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(234): here
  3731. function "__half::operator unsigned long long() const"
  3732. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(237): here
  3733. function "__half::operator __nv_bool() const"
  3734. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(241): here
  3735. detected during:
  3736. instantiation of "void tcnn::warp_activation_backward_in<T,fragment_t,forward_fragment_t>(tcnn::Activatio
  3737. n, const fragment_t &, const forward_fragment_t &, fragment_t &) [with T=tcnn::network_precision_t, fragment_t=tcnn::
  3738. VectorFragment<tcnn::vector_t<tcnn::network_precision_t, 8U>>, forward_fragment_t=tcnn::VectorFragment<tcnn::vector_t
  3739. <tcnn::network_precision_t, 8U>>]"
  3740. (256): here
  3741. instantiation of "void tcnn::kernel_activation_backward(uint32_t, tcnn::Activation, const T *, const T *,
  3742. T *) [with T=tcnn::network_precision_t, N=8U]"
  3743. (310): here
  3744. instantiation of "void tcnn::activation_backward_gpu(cudaStream_t, uint32_t, tcnn::Activation, const T *,
  3745. const T *, T *) [with T=tcnn::network_precision_t]"
  3746. (319): here
  3747. instantiation of "void tcnn::activation_backward_gpu(cudaStream_t, tcnn::Activation, const tcnn::GPUMatri
  3748. xDynamic<T> &, tcnn::GPUMatrixDynamic<T> &) [with T=tcnn::network_precision_t]"
  3749. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_resnet.cu(323): here
  3750. instantiation of "void tcnn::CutlassResNet<T, input_activation>::backward(cudaStream_t, const tcnn::Conte
  3751. xt &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, tcnn::
  3752. GPUMatrixDynamic<T> *, __nv_bool, __nv_bool) [with T=tcnn::network_precision_t, input_activation=tcnn::Activation::No
  3753. ne]"
  3754. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_resnet.cu(391): here
  3755.  
  3756. C:\Program Files (x86)\Microsoft Visual Studio\2019\Community\MSBuild\Microsoft\VC\v160\BuildCustomizations\CUDA 11.6.t
  3757. argets(790,9): error MSB3721: The command ""C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\bin\nvcc.exe" -gen
  3758. code=arch=compute_52,code=\"sm_52,compute_52\" --use-local-env -ccbin "C:\Program Files (x86)\Microsoft Visual Studio\2
  3759. 019\Community\VC\Tools\MSVC\14.29.30133\bin\HostX64\x64" -x cu -I"C:\ngp\instant-ngp\dependencies" -I"C:\ngp\instant-
  3760. ngp\dependencies\eigen" -I"C:\ngp\instant-ngp\dependencies\filesystem" -I"C:\ngp\instant-ngp\dependencies\glfw\include"
  3761. -I"C:\ngp\instant-ngp\dependencies\imgui\gl3w" -I"C:\ngp\instant-ngp\dependencies\nanovdb" -I"C:\ProgramData\NVIDIA Co
  3762. rporation\OptiX SDK 7.4.0\include" -I"C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include" -I"C:\ngp\instant-ngp\depen
  3763. dencies\tiny-cuda-nn\dependencies" -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include" --keep-dir
  3764. x64\RelWithDebInfo -maxrregcount=0 --machine 64 --compile -cudart static --extended-lambda --expt-relaxed-constexpr -
  3765. std=c++14 -Xcompiler="/EHsc -Zi -Ob1 -bigobj" -D_WINDOWS -DNDEBUG -DNGP_GUI -DNGP_OPTIX -DTCNN_MIN_GPU_ARCH=75 -DTCNN
  3766. _SHAMPOO -D"CMAKE_INTDIR=\"RelWithDebInfo\"" -D_MBCS -D"CMAKE_INTDIR=\"RelWithDebInfo\"" -Xcompiler "/EHsc /W1 /nologo
  3767. /O2 /FdC:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\RelWithDebInfo\tiny-cuda-nn.pdb /FS /Zi /MD /GR" -o tiny
  3768. -cuda-nn.dir\RelWithDebInfo\cutlass_mlp.obj "C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_mlp.cu"" exited w
  3769. ith code 1. [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  3770. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/common_device.h(148): error : more than one conversio
  3771. n function from "tcnn::network_precision_t" to a built-in type applies: [C:\ngp\instant-ngp\build\dependencies\tiny-cud
  3772. a-nn\src\tiny-cuda-nn.vcxproj]
  3773. function "__half::operator float() const"
  3774. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(204): here
  3775. function "__half::operator short() const"
  3776. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(222): here
  3777. function "__half::operator unsigned short() const"
  3778. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(225): here
  3779. function "__half::operator int() const"
  3780. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(228): here
  3781. function "__half::operator unsigned int() const"
  3782. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(231): here
  3783. function "__half::operator long long() const"
  3784. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(234): here
  3785. function "__half::operator unsigned long long() const"
  3786. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(237): here
  3787. function "__half::operator __nv_bool() const"
  3788. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(241): here
  3789. detected during:
  3790. instantiation of "void tcnn::warp_activation_backward_in<T,fragment_t,forward_fragment_t>(tcnn::Activatio
  3791. n, const fragment_t &, const forward_fragment_t &, fragment_t &) [with T=tcnn::network_precision_t, fragment_t=tcnn::
  3792. VectorFragment<tcnn::vector_t<tcnn::network_precision_t, 8U>>, forward_fragment_t=tcnn::VectorFragment<tcnn::vector_t
  3793. <tcnn::network_precision_t, 8U>>]"
  3794. (256): here
  3795. instantiation of "void tcnn::kernel_activation_backward(uint32_t, tcnn::Activation, const T *, const T *,
  3796. T *) [with T=tcnn::network_precision_t, N=8U]"
  3797. (310): here
  3798. instantiation of "void tcnn::activation_backward_gpu(cudaStream_t, uint32_t, tcnn::Activation, const T *,
  3799. const T *, T *) [with T=tcnn::network_precision_t]"
  3800. (319): here
  3801. instantiation of "void tcnn::activation_backward_gpu(cudaStream_t, tcnn::Activation, const tcnn::GPUMatri
  3802. xDynamic<T> &, tcnn::GPUMatrixDynamic<T> &) [with T=tcnn::network_precision_t]"
  3803. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_resnet.cu(323): here
  3804. instantiation of "void tcnn::CutlassResNet<T, input_activation>::backward(cudaStream_t, const tcnn::Conte
  3805. xt &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, tcnn::
  3806. GPUMatrixDynamic<T> *, __nv_bool, __nv_bool) [with T=tcnn::network_precision_t, input_activation=tcnn::Activation::No
  3807. ne]"
  3808. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_resnet.cu(391): here
  3809.  
  3810. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/common_device.h(148): error : more than one conversio
  3811. n function from "tcnn::network_precision_t" to a built-in type applies: [C:\ngp\instant-ngp\build\dependencies\tiny-cud
  3812. a-nn\src\tiny-cuda-nn.vcxproj]
  3813. function "__half::operator float() const"
  3814. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(204): here
  3815. function "__half::operator short() const"
  3816. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(222): here
  3817. function "__half::operator unsigned short() const"
  3818. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(225): here
  3819. function "__half::operator int() const"
  3820. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(228): here
  3821. function "__half::operator unsigned int() const"
  3822. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(231): here
  3823. function "__half::operator long long() const"
  3824. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(234): here
  3825. function "__half::operator unsigned long long() const"
  3826. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(237): here
  3827. function "__half::operator __nv_bool() const"
  3828. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(241): here
  3829. detected during:
  3830. instantiation of "void tcnn::warp_activation_backward_in<T,fragment_t,forward_fragment_t>(tcnn::Activatio
  3831. n, const fragment_t &, const forward_fragment_t &, fragment_t &) [with T=tcnn::network_precision_t, fragment_t=tcnn::
  3832. VectorFragment<tcnn::vector_t<tcnn::network_precision_t, 8U>>, forward_fragment_t=tcnn::VectorFragment<tcnn::vector_t
  3833. <tcnn::network_precision_t, 8U>>]"
  3834. (256): here
  3835. instantiation of "void tcnn::kernel_activation_backward(uint32_t, tcnn::Activation, const T *, const T *,
  3836. T *) [with T=tcnn::network_precision_t, N=8U]"
  3837. (310): here
  3838. instantiation of "void tcnn::activation_backward_gpu(cudaStream_t, uint32_t, tcnn::Activation, const T *,
  3839. const T *, T *) [with T=tcnn::network_precision_t]"
  3840. (319): here
  3841. instantiation of "void tcnn::activation_backward_gpu(cudaStream_t, tcnn::Activation, const tcnn::GPUMatri
  3842. xDynamic<T> &, tcnn::GPUMatrixDynamic<T> &) [with T=tcnn::network_precision_t]"
  3843. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_resnet.cu(323): here
  3844. instantiation of "void tcnn::CutlassResNet<T, input_activation>::backward(cudaStream_t, const tcnn::Conte
  3845. xt &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, tcnn::
  3846. GPUMatrixDynamic<T> *, __nv_bool, __nv_bool) [with T=tcnn::network_precision_t, input_activation=tcnn::Activation::No
  3847. ne]"
  3848. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_resnet.cu(391): here
  3849.  
  3850. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/common_device.h(156): error : more than one conversio
  3851. n function from "tcnn::network_precision_t" to a built-in type applies: [C:\ngp\instant-ngp\build\dependencies\tiny-cud
  3852. a-nn\src\tiny-cuda-nn.vcxproj]
  3853. function "__half::operator float() const"
  3854. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(204): here
  3855. function "__half::operator short() const"
  3856. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(222): here
  3857. function "__half::operator unsigned short() const"
  3858. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(225): here
  3859. function "__half::operator int() const"
  3860. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(228): here
  3861. function "__half::operator unsigned int() const"
  3862. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(231): here
  3863. function "__half::operator long long() const"
  3864. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(234): here
  3865. function "__half::operator unsigned long long() const"
  3866. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(237): here
  3867. function "__half::operator __nv_bool() const"
  3868. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(241): here
  3869. detected during:
  3870. instantiation of "void tcnn::warp_activation_backward_in<T,fragment_t,forward_fragment_t>(tcnn::Activatio
  3871. n, const fragment_t &, const forward_fragment_t &, fragment_t &) [with T=tcnn::network_precision_t, fragment_t=tcnn::
  3872. VectorFragment<tcnn::vector_t<tcnn::network_precision_t, 8U>>, forward_fragment_t=tcnn::VectorFragment<tcnn::vector_t
  3873. <tcnn::network_precision_t, 8U>>]"
  3874. (256): here
  3875. instantiation of "void tcnn::kernel_activation_backward(uint32_t, tcnn::Activation, const T *, const T *,
  3876. T *) [with T=tcnn::network_precision_t, N=8U]"
  3877. (310): here
  3878. instantiation of "void tcnn::activation_backward_gpu(cudaStream_t, uint32_t, tcnn::Activation, const T *,
  3879. const T *, T *) [with T=tcnn::network_precision_t]"
  3880. (319): here
  3881. instantiation of "void tcnn::activation_backward_gpu(cudaStream_t, tcnn::Activation, const tcnn::GPUMatri
  3882. xDynamic<T> &, tcnn::GPUMatrixDynamic<T> &) [with T=tcnn::network_precision_t]"
  3883. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_resnet.cu(323): here
  3884. instantiation of "void tcnn::CutlassResNet<T, input_activation>::backward(cudaStream_t, const tcnn::Conte
  3885. xt &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, tcnn::
  3886. GPUMatrixDynamic<T> *, __nv_bool, __nv_bool) [with T=tcnn::network_precision_t, input_activation=tcnn::Activation::No
  3887. ne]"
  3888. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_resnet.cu(391): here
  3889.  
  3890. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/common_device.h(156): error : more than one conversio
  3891. n function from "tcnn::network_precision_t" to a built-in type applies: [C:\ngp\instant-ngp\build\dependencies\tiny-cud
  3892. a-nn\src\tiny-cuda-nn.vcxproj]
  3893. function "__half::operator float() const"
  3894. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(204): here
  3895. function "__half::operator short() const"
  3896. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(222): here
  3897. function "__half::operator unsigned short() const"
  3898. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(225): here
  3899. function "__half::operator int() const"
  3900. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(228): here
  3901. function "__half::operator unsigned int() const"
  3902. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(231): here
  3903. function "__half::operator long long() const"
  3904. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(234): here
  3905. function "__half::operator unsigned long long() const"
  3906. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(237): here
  3907. function "__half::operator __nv_bool() const"
  3908. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(241): here
  3909. detected during:
  3910. instantiation of "void tcnn::warp_activation_backward_in<T,fragment_t,forward_fragment_t>(tcnn::Activatio
  3911. n, const fragment_t &, const forward_fragment_t &, fragment_t &) [with T=tcnn::network_precision_t, fragment_t=tcnn::
  3912. VectorFragment<tcnn::vector_t<tcnn::network_precision_t, 8U>>, forward_fragment_t=tcnn::VectorFragment<tcnn::vector_t
  3913. <tcnn::network_precision_t, 8U>>]"
  3914. (256): here
  3915. instantiation of "void tcnn::kernel_activation_backward(uint32_t, tcnn::Activation, const T *, const T *,
  3916. T *) [with T=tcnn::network_precision_t, N=8U]"
  3917. (310): here
  3918. instantiation of "void tcnn::activation_backward_gpu(cudaStream_t, uint32_t, tcnn::Activation, const T *,
  3919. const T *, T *) [with T=tcnn::network_precision_t]"
  3920. (319): here
  3921. instantiation of "void tcnn::activation_backward_gpu(cudaStream_t, tcnn::Activation, const tcnn::GPUMatri
  3922. xDynamic<T> &, tcnn::GPUMatrixDynamic<T> &) [with T=tcnn::network_precision_t]"
  3923. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_resnet.cu(323): here
  3924. instantiation of "void tcnn::CutlassResNet<T, input_activation>::backward(cudaStream_t, const tcnn::Conte
  3925. xt &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, tcnn::
  3926. GPUMatrixDynamic<T> *, __nv_bool, __nv_bool) [with T=tcnn::network_precision_t, input_activation=tcnn::Activation::No
  3927. ne]"
  3928. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_resnet.cu(391): here
  3929.  
  3930. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/common_device.h(163): error : more than one conversio
  3931. n function from "tcnn::network_precision_t" to a built-in type applies: [C:\ngp\instant-ngp\build\dependencies\tiny-cud
  3932. a-nn\src\tiny-cuda-nn.vcxproj]
  3933. function "__half::operator float() const"
  3934. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(204): here
  3935. function "__half::operator short() const"
  3936. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(222): here
  3937. function "__half::operator unsigned short() const"
  3938. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(225): here
  3939. function "__half::operator int() const"
  3940. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(228): here
  3941. function "__half::operator unsigned int() const"
  3942. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(231): here
  3943. function "__half::operator long long() const"
  3944. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(234): here
  3945. function "__half::operator unsigned long long() const"
  3946. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(237): here
  3947. function "__half::operator __nv_bool() const"
  3948. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(241): here
  3949. detected during:
  3950. instantiation of "void tcnn::warp_activation_backward_in<T,fragment_t,forward_fragment_t>(tcnn::Activatio
  3951. n, const fragment_t &, const forward_fragment_t &, fragment_t &) [with T=tcnn::network_precision_t, fragment_t=tcnn::
  3952. VectorFragment<tcnn::vector_t<tcnn::network_precision_t, 8U>>, forward_fragment_t=tcnn::VectorFragment<tcnn::vector_t
  3953. <tcnn::network_precision_t, 8U>>]"
  3954. (256): here
  3955. instantiation of "void tcnn::kernel_activation_backward(uint32_t, tcnn::Activation, const T *, const T *,
  3956. T *) [with T=tcnn::network_precision_t, N=8U]"
  3957. (310): here
  3958. instantiation of "void tcnn::activation_backward_gpu(cudaStream_t, uint32_t, tcnn::Activation, const T *,
  3959. const T *, T *) [with T=tcnn::network_precision_t]"
  3960. (319): here
  3961. instantiation of "void tcnn::activation_backward_gpu(cudaStream_t, tcnn::Activation, const tcnn::GPUMatri
  3962. xDynamic<T> &, tcnn::GPUMatrixDynamic<T> &) [with T=tcnn::network_precision_t]"
  3963. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_resnet.cu(323): here
  3964. instantiation of "void tcnn::CutlassResNet<T, input_activation>::backward(cudaStream_t, const tcnn::Conte
  3965. xt &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, tcnn::
  3966. GPUMatrixDynamic<T> *, __nv_bool, __nv_bool) [with T=tcnn::network_precision_t, input_activation=tcnn::Activation::No
  3967. ne]"
  3968. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_resnet.cu(391): here
  3969.  
  3970. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include\tiny-cuda-nn/common_device.h(163): error : more than one conversio
  3971. n function from "tcnn::network_precision_t" to a built-in type applies: [C:\ngp\instant-ngp\build\dependencies\tiny-cud
  3972. a-nn\src\tiny-cuda-nn.vcxproj]
  3973. function "__half::operator float() const"
  3974. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(204): here
  3975. function "__half::operator short() const"
  3976. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(222): here
  3977. function "__half::operator unsigned short() const"
  3978. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(225): here
  3979. function "__half::operator int() const"
  3980. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(228): here
  3981. function "__half::operator unsigned int() const"
  3982. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(231): here
  3983. function "__half::operator long long() const"
  3984. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(234): here
  3985. function "__half::operator unsigned long long() const"
  3986. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(237): here
  3987. function "__half::operator __nv_bool() const"
  3988. C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include\cuda_fp16.hpp(241): here
  3989. detected during:
  3990. instantiation of "void tcnn::warp_activation_backward_in<T,fragment_t,forward_fragment_t>(tcnn::Activatio
  3991. n, const fragment_t &, const forward_fragment_t &, fragment_t &) [with T=tcnn::network_precision_t, fragment_t=tcnn::
  3992. VectorFragment<tcnn::vector_t<tcnn::network_precision_t, 8U>>, forward_fragment_t=tcnn::VectorFragment<tcnn::vector_t
  3993. <tcnn::network_precision_t, 8U>>]"
  3994. (256): here
  3995. instantiation of "void tcnn::kernel_activation_backward(uint32_t, tcnn::Activation, const T *, const T *,
  3996. T *) [with T=tcnn::network_precision_t, N=8U]"
  3997. (310): here
  3998. instantiation of "void tcnn::activation_backward_gpu(cudaStream_t, uint32_t, tcnn::Activation, const T *,
  3999. const T *, T *) [with T=tcnn::network_precision_t]"
  4000. (319): here
  4001. instantiation of "void tcnn::activation_backward_gpu(cudaStream_t, tcnn::Activation, const tcnn::GPUMatri
  4002. xDynamic<T> &, tcnn::GPUMatrixDynamic<T> &) [with T=tcnn::network_precision_t]"
  4003. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_resnet.cu(323): here
  4004. instantiation of "void tcnn::CutlassResNet<T, input_activation>::backward(cudaStream_t, const tcnn::Conte
  4005. xt &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, const tcnn::GPUMatrixDynamic<T> &, tcnn::
  4006. GPUMatrixDynamic<T> *, __nv_bool, __nv_bool) [with T=tcnn::network_precision_t, input_activation=tcnn::Activation::No
  4007. ne]"
  4008. C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_resnet.cu(391): here
  4009.  
  4010. 26 errors detected in the compilation of "C:/ngp/instant-ngp/dependencies/tiny-cuda-nn/src/cutlass_resnet.cu".
  4011. cutlass_resnet.cu
  4012. C:\Program Files (x86)\Microsoft Visual Studio\2019\Community\MSBuild\Microsoft\VC\v160\BuildCustomizations\CUDA 11.6.t
  4013. argets(790,9): error MSB3721: The command ""C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\bin\nvcc.exe" -gen
  4014. code=arch=compute_52,code=\"sm_52,compute_52\" --use-local-env -ccbin "C:\Program Files (x86)\Microsoft Visual Studio\2
  4015. 019\Community\VC\Tools\MSVC\14.29.30133\bin\HostX64\x64" -x cu -I"C:\ngp\instant-ngp\dependencies" -I"C:\ngp\instant-
  4016. ngp\dependencies\eigen" -I"C:\ngp\instant-ngp\dependencies\filesystem" -I"C:\ngp\instant-ngp\dependencies\glfw\include"
  4017. -I"C:\ngp\instant-ngp\dependencies\imgui\gl3w" -I"C:\ngp\instant-ngp\dependencies\nanovdb" -I"C:\ProgramData\NVIDIA Co
  4018. rporation\OptiX SDK 7.4.0\include" -I"C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\include" -I"C:\ngp\instant-ngp\depen
  4019. dencies\tiny-cuda-nn\dependencies" -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\include" --keep-dir
  4020. x64\RelWithDebInfo -maxrregcount=0 --machine 64 --compile -cudart static --extended-lambda --expt-relaxed-constexpr -
  4021. std=c++14 -Xcompiler="/EHsc -Zi -Ob1 -bigobj" -D_WINDOWS -DNDEBUG -DNGP_GUI -DNGP_OPTIX -DTCNN_MIN_GPU_ARCH=75 -DTCNN
  4022. _SHAMPOO -D"CMAKE_INTDIR=\"RelWithDebInfo\"" -D_MBCS -D"CMAKE_INTDIR=\"RelWithDebInfo\"" -Xcompiler "/EHsc /W1 /nologo
  4023. /O2 /FdC:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\RelWithDebInfo\tiny-cuda-nn.pdb /FS /Zi /MD /GR" -o tiny
  4024. -cuda-nn.dir\RelWithDebInfo\cutlass_resnet.obj "C:\ngp\instant-ngp\dependencies\tiny-cuda-nn\src\cutlass_resnet.cu"" ex
  4025. ited with code 1. [C:\ngp\instant-ngp\build\dependencies\tiny-cuda-nn\src\tiny-cuda-nn.vcxproj]
  4026. common.cu
  4027. reduce_sum.cu
  4028. common_device.cu
  4029. object.cu
  4030. network.cu
  4031. cpp_api.cu
  4032. loss.cu
  4033. optimizer.cu
  4034.  
  4035. C:\ngp\instant-ngp>
Advertisement
Add Comment
Please, Sign In to add comment
Advertisement
Public Pastes
Advertisement
We use cookies for various purposes including analytics. By continuing to use Pastebin, you agree to our use of cookies as described in the Cookies Policy.  OK, I Understand
Not a member of Pastebin yet?
Sign Up, it unlocks many cool features!