API tools faq
paste
Advertisement
Guest User

Untitled

a guest
Jul 21st, 2017
53
0
Never
Add comment
Not a member of Pastebin yet? Sign Up, it unlocks many cool features!
text 10.67 KB | None | 0 0
raw download clone embed print report
  1. __attribute__((noinline))
  2. void benchmark(cfloat* __restrict__ aa, cfloat* __restrict__ bb, cfloat* __restrict__ cc, cfloat* __restrict__ dd, cfloat uu, cfloat vv, size_t nn) {
  3. for (ssize_t ii=0; ii < nn; ii++) {
  4. dd[ii] = (
  5. aa[ii]*uu +
  6. bb[ii]*vv +
  7. cc[ii]
  8. );
  9. }
  10. }
  11.  
  12. struct cfloat {
  13. cfloat(float re, float im) : re(re), im(im) {}
  14. float re,im;
  15. };
  16.  
  17. cfloat operator +(cfloat a, cfloat b) {
  18. return cfloat(a.re+b.re, a.im+b.im);
  19. }
  20.  
  21. cfloat operator *(cfloat a, cfloat b) {
  22. return cfloat(a.re*b.re-a.im*b.im, a.re*b.im+a.im*b.re);
  23. }
  24.  
  25. 0x00000000004006a0 <+0>: push %r15
  26. 0x00000000004006a2 <+2>: test %r8,%r8
  27. 0x00000000004006a5 <+5>: push %r14
  28. 0x00000000004006a7 <+7>: push %r13
  29. 0x00000000004006a9 <+9>: push %r12
  30. 0x00000000004006ab <+11>: push %rbp
  31. 0x00000000004006ac <+12>: push %rbx
  32. 0x00000000004006ad <+13>: movq %xmm0,-0x28(%rsp)
  33. 0x00000000004006b3 <+19>: mov %rdi,-0x38(%rsp)
  34. 0x00000000004006b8 <+24>: mov -0x28(%rsp),%rax
  35. 0x00000000004006bd <+29>: movq %xmm1,-0x28(%rsp)
  36. 0x00000000004006c3 <+35>: mov -0x28(%rsp),%r9
  37. 0x00000000004006c8 <+40>: je 0x4008a0 <_Z9benchmarkP6cfloatS0_S0_S0_S_S_m+512>
  38. 0x00000000004006ce <+46>: mov %r9,%r15
  39. 0x00000000004006d1 <+49>: mov %rax,%r14
  40. 0x00000000004006d4 <+52>: xor %r11d,%r11d
  41. 0x00000000004006d7 <+55>: shr $0x20,%r15
  42. 0x00000000004006db <+59>: shr $0x20,%r14
  43. 0x00000000004006df <+63>: xor %r10d,%r10d
  44. 0x00000000004006e2 <+66>: mov %r15d,-0x2c(%rsp)
  45. 0x00000000004006e7 <+71>: xor %ebp,%ebp
  46. 0x00000000004006e9 <+73>: xor %ebx,%ebx
  47. 0x00000000004006eb <+75>: movss -0x2c(%rsp),%xmm6
  48. 0x00000000004006f1 <+81>: mov %r9d,-0x2c(%rsp)
  49. 0x00000000004006f6 <+86>: movss -0x2c(%rsp),%xmm5
  50. 0x00000000004006fc <+92>: mov %r14d,-0x2c(%rsp)
  51. 0x0000000000400701 <+97>: movss -0x2c(%rsp),%xmm4
  52. 0x0000000000400707 <+103>: mov %eax,-0x2c(%rsp)
  53. 0x000000000040070b <+107>: xor %r13d,%r13d
  54. 0x000000000040070e <+110>: xor %r12d,%r12d
  55. 0x0000000000400711 <+113>: movabs $0xffffffff00000000,%r9
  56. 0x000000000040071b <+123>: movss -0x2c(%rsp),%xmm3
  57. 0x0000000000400721 <+129>: nopl 0x0(%rax)
  58. 0x0000000000400728 <+136>: lea 0x0(,%r13,8),%rax
  59. 0x0000000000400730 <+144>: movaps %xmm6,%xmm1
  60. 0x0000000000400733 <+147>: movaps %xmm5,%xmm7
  61. 0x0000000000400736 <+150>: and $0xffffffff,%ebp
  62. 0x0000000000400739 <+153>: lea (%rsi,%rax,1),%r15
  63. 0x000000000040073d <+157>: lea (%rdx,%rax,1),%r14
  64. 0x0000000000400741 <+161>: add -0x38(%rsp),%rax
  65. 0x0000000000400746 <+166>: and $0xffffffff,%ebx
  66. 0x0000000000400749 <+169>: add $0x1,%r12
  67. 0x000000000040074d <+173>: movss (%r15),%xmm0
  68. 0x0000000000400752 <+178>: movss 0x4(%r15),%xmm2
  69. 0x0000000000400758 <+184>: mulss %xmm0,%xmm1
  70. 0x000000000040075c <+188>: mulss %xmm2,%xmm7
  71. 0x0000000000400760 <+192>: mulss %xmm5,%xmm0
  72. 0x0000000000400764 <+196>: mulss %xmm6,%xmm2
  73. 0x0000000000400768 <+200>: addss %xmm7,%xmm1
  74. 0x000000000040076c <+204>: movaps %xmm3,%xmm7
  75. 0x000000000040076f <+207>: subss %xmm2,%xmm0
  76. 0x0000000000400773 <+211>: movd %xmm1,-0x30(%rsp)
  77. 0x0000000000400779 <+217>: mov -0x30(%rsp),%edi
  78. 0x000000000040077d <+221>: movaps %xmm4,%xmm1
  79. 0x0000000000400780 <+224>: movd %xmm0,-0x30(%rsp)
  80. 0x0000000000400786 <+230>: mov %edi,%r15d
  81. 0x0000000000400789 <+233>: mov -0x30(%rsp),%edi
  82. 0x000000000040078d <+237>: movss (%rax),%xmm0
  83. 0x0000000000400791 <+241>: shl $0x20,%r15
  84. 0x0000000000400795 <+245>: movss 0x4(%rax),%xmm2
  85. 0x000000000040079a <+250>: mulss %xmm0,%xmm1
  86. 0x000000000040079e <+254>: or %r15,%rbp
  87. 0x00000000004007a1 <+257>: mulss %xmm2,%xmm7
  88. 0x00000000004007a5 <+261>: mov %edi,%r15d
  89. 0x00000000004007a8 <+264>: and %r9,%rbp
  90. 0x00000000004007ab <+267>: mulss %xmm3,%xmm0
  91. 0x00000000004007af <+271>: or %r15,%rbp
  92. 0x00000000004007b2 <+274>: mulss %xmm4,%xmm2
  93. 0x00000000004007b6 <+278>: addss %xmm7,%xmm1
  94. 0x00000000004007ba <+282>: subss %xmm2,%xmm0
  95. 0x00000000004007be <+286>: movd %xmm1,-0x30(%rsp)
  96. 0x00000000004007c4 <+292>: mov -0x30(%rsp),%edi
  97. 0x00000000004007c8 <+296>: movd %xmm0,-0x30(%rsp)
  98. 0x00000000004007ce <+302>: mov %edi,%eax
  99. 0x00000000004007d0 <+304>: mov -0x30(%rsp),%edi
  100. 0x00000000004007d4 <+308>: shl $0x20,%rax
  101. 0x00000000004007d8 <+312>: or %rax,%rbx
  102. 0x00000000004007db <+315>: and %r9,%rbx
  103. 0x00000000004007de <+318>: mov %edi,%eax
  104. 0x00000000004007e0 <+320>: or %rax,%rbx
  105. 0x00000000004007e3 <+323>: mov %r10,%rax
  106. 0x00000000004007e6 <+326>: mov %rbx,%rdi
  107. 0x00000000004007e9 <+329>: and $0xffffffff,%eax
  108. 0x00000000004007ec <+332>: shr $0x20,%rdi
  109. 0x00000000004007f0 <+336>: mov %edi,-0x20(%rsp)
  110. 0x00000000004007f4 <+340>: mov %rbp,%rdi
  111. 0x00000000004007f7 <+343>: shr $0x20,%rdi
  112. 0x00000000004007fb <+347>: movss -0x20(%rsp),%xmm0
  113. 0x0000000000400801 <+353>: mov %edi,-0x10(%rsp)
  114. 0x0000000000400805 <+357>: addss -0x10(%rsp),%xmm0
  115. 0x000000000040080b <+363>: mov %ebp,-0x10(%rsp)
  116. 0x000000000040080f <+367>: movss %xmm0,-0x20(%rsp)
  117. 0x0000000000400815 <+373>: mov -0x20(%rsp),%r10d
  118. 0x000000000040081a <+378>: mov %ebx,-0x20(%rsp)
  119. 0x000000000040081e <+382>: movss -0x20(%rsp),%xmm0
  120. 0x0000000000400824 <+388>: addss -0x10(%rsp),%xmm0
  121. 0x000000000040082a <+394>: shl $0x20,%r10
  122. 0x000000000040082e <+398>: or %rax,%r10
  123. 0x0000000000400831 <+401>: and %r9,%r10
  124. 0x0000000000400834 <+404>: movss %xmm0,-0x20(%rsp)
  125. 0x000000000040083a <+410>: mov -0x20(%rsp),%eax
  126. 0x000000000040083e <+414>: or %rax,%r10
  127. 0x0000000000400841 <+417>: mov %r11,%rax
  128. 0x0000000000400844 <+420>: mov %r10,%rdi
  129. 0x0000000000400847 <+423>: and $0xffffffff,%eax
  130. 0x000000000040084a <+426>: shr $0x20,%rdi
  131. 0x000000000040084e <+430>: mov %edi,-0x20(%rsp)
  132. 0x0000000000400852 <+434>: movss -0x20(%rsp),%xmm0
  133. 0x0000000000400858 <+440>: addss 0x4(%r14),%xmm0
  134. 0x000000000040085e <+446>: movss %xmm0,-0x20(%rsp)
  135. 0x0000000000400864 <+452>: mov -0x20(%rsp),%r11d
  136. 0x0000000000400869 <+457>: mov %r10d,-0x20(%rsp)
  137. 0x000000000040086e <+462>: movss -0x20(%rsp),%xmm0
  138. 0x0000000000400874 <+468>: addss (%r14),%xmm0
  139. 0x0000000000400879 <+473>: shl $0x20,%r11
  140. 0x000000000040087d <+477>: or %rax,%r11
  141. 0x0000000000400880 <+480>: and %r9,%r11
  142. 0x0000000000400883 <+483>: movss %xmm0,-0x20(%rsp)
  143. 0x0000000000400889 <+489>: mov -0x20(%rsp),%eax
  144. 0x000000000040088d <+493>: or %rax,%r11
  145. 0x0000000000400890 <+496>: cmp %r8,%r12
  146. 0x0000000000400893 <+499>: mov %r11,(%rcx,%r13,8)
  147. 0x0000000000400897 <+503>: mov %r12,%r13
  148. 0x000000000040089a <+506>: jne 0x400728 <_Z9benchmarkP6cfloatS0_S0_S0_S_S_m+136>
  149. 0x00000000004008a0 <+512>: pop %rbx
  150. 0x00000000004008a1 <+513>: pop %rbp
  151. 0x00000000004008a2 <+514>: pop %r12
  152. 0x00000000004008a4 <+516>: pop %r13
  153. 0x00000000004008a6 <+518>: pop %r14
  154. 0x00000000004008a8 <+520>: pop %r15
  155. 0x00000000004008aa <+522>: retq
  156.  
  157. struct cfloat {
  158. cfloat(float re, float im) { ri[0] = re; ri[1] = im; }
  159. float ri[2];
  160. };
  161.  
  162. cfloat operator +(cfloat a, cfloat b) {
  163. return cfloat(a.ri[0]+b.ri[0], a.ri[1]+b.ri[1]);
  164. }
  165.  
  166. cfloat operator *(cfloat a, cfloat b) {
  167. return cfloat(a.ri[0]*b.ri[0]-a.ri[1]*b.ri[1], a.ri[0]*b.ri[1]+a.ri[1]*b.ri[0]);
  168. }
  169.  
  170. Dump of assembler code for function _Z9benchmarkP6cfloatS0_S0_S0_S_S_m:
  171. 0x00000000004006a0 <+0>: push %rbx
  172. 0x00000000004006a1 <+1>: movq %xmm0,-0x8(%rsp)
  173. 0x00000000004006a7 <+7>: mov -0x8(%rsp),%r9
  174. 0x00000000004006ac <+12>: movq %xmm1,-0x8(%rsp)
  175. 0x00000000004006b2 <+18>: mov -0x8(%rsp),%rax
  176. 0x00000000004006b7 <+23>: mov %r9d,-0xc(%rsp)
  177. 0x00000000004006bc <+28>: shr $0x20,%r9
  178. 0x00000000004006c0 <+32>: movss -0xc(%rsp),%xmm9
  179. 0x00000000004006c7 <+39>: mov %r9d,-0xc(%rsp)
  180. 0x00000000004006cc <+44>: movss -0xc(%rsp),%xmm8
  181. 0x00000000004006d3 <+51>: mov %eax,-0xc(%rsp)
  182. 0x00000000004006d7 <+55>: shr $0x20,%rax
  183. 0x00000000004006db <+59>: movss -0xc(%rsp),%xmm7
  184. 0x00000000004006e1 <+65>: test %r8,%r8
  185. 0x00000000004006e4 <+68>: mov %eax,-0xc(%rsp)
  186. 0x00000000004006e8 <+72>: movss -0xc(%rsp),%xmm6
  187. 0x00000000004006ee <+78>: je 0x400796 <_Z9benchmarkP6cfloatS0_S0_S0_S_S_m+246>
  188. 0x00000000004006f4 <+84>: xor %eax,%eax
  189. 0x00000000004006f6 <+86>: xor %r9d,%r9d
  190. 0x00000000004006f9 <+89>: nopl 0x0(%rax)
  191. 0x0000000000400700 <+96>: shl $0x3,%rax
  192. 0x0000000000400704 <+100>: movaps %xmm7,%xmm0
  193. 0x0000000000400707 <+103>: lea (%rsi,%rax,1),%rbx
  194. 0x000000000040070b <+107>: movaps %xmm6,%xmm3
  195. 0x000000000040070e <+110>: lea (%rcx,%rax,1),%r10
  196. 0x0000000000400712 <+114>: lea (%rdx,%rax,1),%r11
  197. 0x0000000000400716 <+118>: lea (%rdi,%rax,1),%rax
  198. 0x000000000040071a <+122>: movss (%rbx),%xmm1
  199. 0x000000000040071e <+126>: add $0x1,%r9
  200. 0x0000000000400722 <+130>: movss 0x4(%rbx),%xmm5
  201. 0x0000000000400727 <+135>: mulss %xmm1,%xmm0
  202. 0x000000000040072b <+139>: mulss %xmm5,%xmm3
  203. 0x000000000040072f <+143>: movss (%rax),%xmm2
  204. 0x0000000000400733 <+147>: movaps %xmm8,%xmm10
  205. 0x0000000000400737 <+151>: mulss %xmm6,%xmm1
  206. 0x000000000040073b <+155>: movss 0x4(%rax),%xmm4
  207. 0x0000000000400740 <+160>: mulss %xmm7,%xmm5
  208. 0x0000000000400744 <+164>: mulss %xmm4,%xmm10
  209. 0x0000000000400749 <+169>: cmp %r8,%r9
  210. 0x000000000040074c <+172>: mov %r9,%rax
  211. 0x000000000040074f <+175>: subss %xmm3,%xmm0
  212. 0x0000000000400753 <+179>: movaps %xmm2,%xmm3
  213. 0x0000000000400756 <+182>: mulss %xmm9,%xmm4
  214. 0x000000000040075b <+187>: mulss %xmm9,%xmm3
  215. 0x0000000000400760 <+192>: addss %xmm5,%xmm1
  216. 0x0000000000400764 <+196>: mulss %xmm8,%xmm2
  217. 0x0000000000400769 <+201>: subss %xmm10,%xmm3
  218. 0x000000000040076e <+206>: addss %xmm4,%xmm2
  219. 0x0000000000400772 <+210>: addss %xmm3,%xmm0
  220. 0x0000000000400776 <+214>: addss %xmm2,%xmm1
  221. 0x000000000040077a <+218>: addss (%r11),%xmm0
  222. 0x000000000040077f <+223>: addss 0x4(%r11),%xmm1
  223. 0x0000000000400785 <+229>: movss %xmm0,(%r10)
  224. 0x000000000040078a <+234>: movss %xmm1,0x4(%r10)
  225. 0x0000000000400790 <+240>: jne 0x400700 <_Z9benchmarkP6cfloatS0_S0_S0_S_S_m+96>
  226. 0x0000000000400796 <+246>: pop %rbx
  227. 0x0000000000400797 <+247>: retq
  228. End of assembler dump.
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!