MD-Bench/arch_analysis/avx512/osaca_force_soa_lt8.txt
Rafael Ravedutti 9c28ff1e9e Add arch_analysis directory and first AVX2 results
Signed-off-by: Rafael Ravedutti <rafaelravedutti@gmail.com>
2021-05-06 23:18:28 +02:00

80 lines
9.3 KiB
Plaintext

iwia021h@testfront1:~/MD-Bench/asm$ /home/hpc/iwia/iwia021h/.local/bin/osaca --ignore-unknown --arch=CSX force_soa_lt8_markers.s
Open Source Architecture Code Analyzer (OSACA) - 0.3.14
Analyzed file: force_soa_lt8_markers.s
Architecture: CSX
Timestamp: 2021-04-29 15:52:48
P - Throughput of LOAD operation can be hidden behind a past or future STORE instruction
* - Instruction micro-ops not bound to a port
X - No throughput/latency information for this instruction in data file
Combined Analysis Report
------------------------
Port pressure in cycles
| 0 - 0DV | 1 | 2 - 2D | 3 - 3D | 4 | 5 | 6 | 7 || CP | LCD |
-------------------------------------------------------------------------------------------------
300 | | | | | | | | || | | # LOE rax rdx rbp rsi rdi r8 r10 r11 r12 ebx r13d xmm8 xmm9 xmm10 ymm17 ymm18 zmm7 zmm11 zmm12 zmm13 zmm14 zmm15 zmm16 zmm19
301 | | | | | | | | || | | ..B1.25: # Preds ..B1.24
302 | | | | | | | | || | | # Execution count [2.50e+01]
303 | | 1.00 | | | | | | || 3.0 | | imulq %r12, %r8 #56.43
304 | | | | | | 1.00 | | || | | vbroadcastsd %xmm9, %zmm9 #59.23
305 | | | | | | 1.00 | | || | 3.0 | vbroadcastsd %xmm8, %zmm2 #58.23
306 | | | | | | 1.00 | | || | | vbroadcastsd %xmm10, %zmm10 #60.23
307 | | | | | | | | || | | X subl %ebx, %r13d #67.9
308 | 0.00 | 0.50 | | | | 0.00 | 0.50 | || 1.0 | | addq %r11, %r8 #37.5
309 | | | | | | | | || | | X vpbroadcastd %r13d, %ymm0 #67.9
310 | | | | | | 1.00 | | || | | vpcmpgtd %ymm17, %ymm0, %k5 #67.9
311 | 0.00 | 0.16 | | | | 0.00 | 0.84 | || | | movslq %ebx, %rbx #67.9
312 | | | | | | | | || | | * vmovaps %zmm19, %zmm4 #70.36
313 | 1.00 | | | | | | | || | | kmovw %k5, %k2 #70.36
314 | | | | | | | | || | | * vmovaps %zmm19, %zmm3 #69.36
315 | 0.00 | 1.00 | 0.50 0.50 | 0.50 0.50 | | 0.00 | | || 0.0 | | vmovdqu32 (%r8,%rbx,4), %ymm1{%k5}{z} #68.21
316 | 1.00 | | | | | | | || | | kmovw %k5, %k1 #69.36
317 | | | | | | | | || | | * vmovaps %zmm19, %zmm5 #71.36
318 | 1.00 | | | | | | | || | | kmovw %k5, %k3 #71.36
319 | 1.50 | 0.50 | 4.00 0.50 | 4.00 0.50 | | 0.50 | 0.50 | || | | vgatherdpd (%rsi,%ymm1,8), %zmm5{%k3} #71.36
320 | 1.50 | 0.34 | 4.00 0.50 | 4.00 0.50 | | 0.50 | 0.66 | || 4.0 | | vgatherdpd (%rax,%ymm1,8), %zmm4{%k2} #70.36
321 | 1.50 | 0.00 | 4.00 0.50 | 4.00 0.50 | | 0.50 | 1.00 | || | | vgatherdpd (%rdx,%ymm1,8), %zmm3{%k1} #69.36
322 | 0.50 | | | | | 0.50 | | || | | vsubpd %zmm5, %zmm10, %zmm30 #71.36
323 | 0.50 | | | | | 0.50 | | || 4.0 | | vsubpd %zmm4, %zmm9, %zmm28 #70.36
324 | 0.50 | | | | | 0.50 | | || | 4.0 | vsubpd %zmm3, %zmm2, %zmm27 #69.36
325 | 0.50 | | | | | 0.50 | | || 4.0 | | vmulpd %zmm28, %zmm28, %zmm26 #72.49
326 | 0.50 | | | | | 0.50 | | || 4.0 | 4.0 | vfmadd231pd %zmm27, %zmm27, %zmm26 #72.49
327 | 0.50 | | | | | 0.50 | | || 4.0 | 4.0 | vfmadd231pd %zmm30, %zmm30, %zmm26 #72.63
328 | 2.50 | | | | | 0.50 | | || 8.0 | 8.0 | vrcp14pd %zmm26, %zmm25 #75.38
329 | | | | | | 1.00 | | || | | vcmppd $1, %zmm16, %zmm26, %k6{%k5} #74.22
330 | | | | | | 1.00 | | || | | vfpclasspd $30, %zmm25, %k0 #75.38
331 | | | | | | | | || | | * vmovaps %zmm26, %zmm6 #75.38
332 | 0.50 | | 0.50 0.50 | 0.50 0.50 | | 0.50 | | || 4.0 | 4.0 | vfnmadd213pd .L_2il0floatpacket.5(%rip){1to8}, %zmm25, %zmm6 #75.38
333 | 1.00 | | | | | | | || | | knotw %k0, %k4 #75.38
334 | 0.50 | | | | | 0.50 | | || 4.0 | 4.0 | vmulpd %zmm6, %zmm6, %zmm8 #75.38
335 | 0.50 | | | | | 0.50 | | || | | vfmadd213pd %zmm25, %zmm6, %zmm25{%k4} #75.38
336 | 0.50 | | | | | 0.50 | | || 4.0 | | vfmadd213pd %zmm25, %zmm8, %zmm25{%k4} #75.38
337 | 0.50 | | | | | 0.50 | | || 4.0 | | vmulpd %zmm15, %zmm25, %zmm20 #76.38
338 | 0.50 | | | | | 0.50 | | || | | vmulpd %zmm14, %zmm25, %zmm22 #77.54
339 | 0.50 | | | | | 0.50 | | || 4.0 | | vmulpd %zmm20, %zmm25, %zmm23 #76.44
340 | 0.50 | | | | | 0.50 | | || 4.0 | | vmulpd %zmm23, %zmm25, %zmm21 #76.50
341 | 0.50 | | | | | 0.50 | | || | | vfmsub213pd %zmm7, %zmm23, %zmm25 #77.54
342 | 0.50 | | | | | 0.50 | | || 4.0 | | vmulpd %zmm22, %zmm21, %zmm24 #77.61
343 | 0.50 | | | | | 0.50 | | || 4.0 | | vmulpd %zmm25, %zmm24, %zmm29 #77.67
344 | 0.00 | | | | | 1.00 | | || | | vfmadd231pd %zmm27, %zmm29, %zmm13{%k6} #78.17
345 | 0.00 | | | | | 1.00 | | || | | vfmadd231pd %zmm28, %zmm29, %zmm12{%k6} #79.17
346 | 0.00 | | | | | 1.00 | | || 4.0 | | vfmadd231pd %zmm30, %zmm29, %zmm11{%k6} #80.17
19.5 3.50 13.0 2.50 13.0 2.50 19.5 3.50 68.0 31.0
Loop-Carried Dependencies Analysis Report
-----------------------------------------
308 | 4.0 | addq %r11, %r8 #37.5| [303, 308]
304 | 3.0 | vbroadcastsd %xmm9, %zmm9 #59.23| [304]
306 | 3.0 | vbroadcastsd %xmm10, %zmm10 #60.23| [306]
307 | 0.0 | subl %ebx, %r13d #67.9| [307]
311 | 1.0 | movslq %ebx, %rbx #67.9| [311]
346 | 4.0 | vfmadd231pd %zmm30, %zmm29, %zmm11{%k6} #80.17| [346]
345 | 4.0 | vfmadd231pd %zmm28, %zmm29, %zmm12{%k6} #79.17| [345]
344 | 4.0 | vfmadd231pd %zmm27, %zmm29, %zmm13{%k6} #78.17| [344]
334 | 31.0 | vmulpd %zmm6, %zmm6, %zmm8 #75.38| [305, 324, 326, 327, 328, 332, 334]