Not a member of Pastebin yet?
Sign Up,
it unlocks many cool features!
- Device 0: NVIDIA GB10
- Compute capability: 12.1
- Total global memory: 119.7 GiB
- Multiprocessor count: 48
- Running benchmarks with target time: 3.0 seconds
- mma_s4s4s32_8_8_32
- run: 3000.8 ms 28.4 T(fl)ops
- mma_mxf4mxf4f32_16_8_64
- run: 2993.9 ms 450.3 T(fl)ops
- mma_nvf4nvf4f32_16_8_64
- run: 2998.7 ms 450.3 T(fl)ops
- mma_f4f4f16_16_8_32
- run: 2997.8 ms 225.3 T(fl)ops
- mma_f4f4f32_16_8_32
- run: 2998.1 ms 225.1 T(fl)ops
- mma_f6f6f16_16_8_32
- run: 2998.3 ms 225.3 T(fl)ops
- mma_f6f6f32_16_8_32
- run: 2998.7 ms 225.1 T(fl)ops
- mma_mxf6mxf6f32_16_8_32
- run: 2998.7 ms 225.1 T(fl)ops
- mma_mxf8mxf8f32_16_8_32
- run: 2998.8 ms 225.1 T(fl)ops
- mma_f8f8f16_16_8_32
- run: 2998.4 ms 225.3 T(fl)ops
- mma_f8f8f32_16_8_32
- run: 2998.7 ms 225.1 T(fl)ops
- mma_s8s8s32_16_16_16
- run: 2993.1 ms 226.6 T(fl)ops
- mma_s8s8s32_32_8_16
- run: 2999.2 ms 226.6 T(fl)ops
- mma_f16f16f16_16_16_16
- run: 2996.9 ms 224.4 T(fl)ops
- mma_f16f16f16_32_8_16
- run: 2998.4 ms 224.4 T(fl)ops
- mma_f16f16f32_16_16_16
- run: 2998.4 ms 224.3 T(fl)ops
- mma_f16f16f32_32_8_16
- run: 2998.4 ms 224.3 T(fl)ops
- mma_bf16bf16f32_16_16_16
- run: 2999.0 ms 224.3 T(fl)ops
- mma_bf16bf16f32_32_8_16
- run: 3014.8 ms 223.2 T(fl)ops
- mma_tf32tf32f32_16_16_8
- run: 2975.9 ms 56.2 T(fl)ops
Advertisement
Add Comment
Please, Sign In to add comment