117 lines
14 KiB
Plaintext
117 lines
14 KiB
Plaintext
|
Open Source Architecture Code Analyzer (OSACA) - 0.4.12
|
||
|
|
Analyzed file: gromacs-icx-avx512-sp.s
|
||
|
|
Architecture: ICX
|
||
|
|
Timestamp: 2023-02-14 12:51:43
|
||
|
|
|
||
|
|
|
||
|
|
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 - 1DV | 2 - 2D | 3 - 3D | 4 | 5 | 6 | 7 | 8 | 9 || CP | LCD |
|
||
|
|
------------------------------------------------------------------------------------------------------------------------
|
||
|
|
1338 | | | | | | | | | | || | | # pointer_increment=64 0f91ac4f7fe1a70d0c899f7f3e745649
|
||
|
|
1339 | | | | | | | | | | || | | # LLVM-MCA-BEGIN
|
||
|
|
1340 | | | | | | | | | | || | | .LBB2_12: # Parent Loop BB2_7 Depth=1
|
||
|
|
1341 | | | | | | | | | | || | | # => This Inner Loop Header: Depth=2
|
||
|
|
1342 | | | 0.50 0.50 | 0.50 0.50 | | | | | | || 5.0 | | movslq (%r11,%rax,4), %rcx
|
||
|
|
1343 | 0.00 | 0.75 | | | | 0.000 | 0.25 | | | || 6.0 | | leaq (%rcx,%rcx,2), %rdx
|
||
|
|
1344 | 0.00 | | | | | | 1.00 | | | || 1.0 | | shlq $5, %rdx
|
||
|
|
1345 | | | 0.50 0.50 | 0.50 0.50 | | | | | | || | | vmovupd (%rsi,%rdx), %zmm16
|
||
|
|
1346 | | | 0.50 0.50 | 0.50 0.50 | | | | | | || 5.0 | | vbroadcastf64x4 64(%rsi,%rdx), %zmm20 # zmm20 = mem[0,1,2,3,0,1,2,3]
|
||
|
|
1347 | | | 0.50 0.50 | 0.50 0.50 | | | | | | || | | vbroadcastf64x4 (%rsi,%rdx), %zmm19 # zmm19 = mem[0,1,2,3,0,1,2,3]
|
||
|
|
1348 | | | | | | 1.000 | | | | || | | vshuff64x2 $238, %zmm16, %zmm16, %zmm21 # zmm21 = zmm16[4,5,6,7,4,5,6,7]
|
||
|
|
1349 | 0.50 | | | | | 0.500 | | | | || | | vsubps %zmm19, %zmm6, %zmm18
|
||
|
|
1350 | 0.50 | | | | | 0.500 | | | | || | | vsubps %zmm21, %zmm10, %zmm17
|
||
|
|
1351 | 0.50 | | | | | 0.500 | | | | || 4.0 | | vsubps %zmm20, %zmm14, %zmm16
|
||
|
|
1352 | 0.50 | | | | | 0.500 | | | | || 4.0 | | vmulps %zmm16, %zmm16, %zmm22
|
||
|
|
1353 | 0.50 | | | | | 0.500 | | | | || 4.0 | | vfmadd231ps %zmm17, %zmm17, %zmm22 # zmm22 = (zmm17 * zmm17) + zmm22
|
||
|
|
1354 | 0.50 | | | | | 0.500 | | | | || 4.0 | | vfmadd231ps %zmm18, %zmm18, %zmm22 # zmm22 = (zmm18 * zmm18) + zmm22
|
||
|
|
1355 | 2.50 | | | | | 0.500 | | | | || 6.0 | | vrcp14ps %zmm22, %zmm23
|
||
|
|
1356 | 0.50 | | | | | 0.500 | | | | || 4.0 | | vmulps %zmm23, %zmm26, %zmm24
|
||
|
|
1357 | 0.50 | | | | | 0.500 | | | | || 4.0 | | vmulps %zmm24, %zmm23, %zmm24
|
||
|
|
1358 | 0.50 | | | | | 0.500 | | | | || 4.0 | | vmulps %zmm24, %zmm23, %zmm24
|
||
|
|
1359 | 0.75 | | | | | 0.250 | | | | || 4.0 | | vaddps %zmm1, %zmm24, %zmm25
|
||
|
|
1360 | 1.00 | | | | | 0.000 | | | | || | | vmulps %zmm23, %zmm27, %zmm23
|
||
|
|
1361 | 1.00 | | | | | 0.000 | | | | || 4.0 | | vmulps %zmm25, %zmm23, %zmm23
|
||
|
|
1362 | 1.00 | | | | | 0.000 | | | | || 4.0 | | vmulps %zmm23, %zmm24, %zmm23
|
||
|
|
1363 | 0.00 | 0.50 | | | | 0.000 | 0.50 | | | || | | leal (%rcx,%rcx), %edx
|
||
|
|
1364 | 0.00 | 0.50 | | | | 0.000 | 0.50 | | | || | | xorl %edi, %edi
|
||
|
|
1365 | 0.00 | 0.50 | | | | 0.000 | 0.50 | | | || | | xorl %ebp, %ebp
|
||
|
|
1366 | 0.00 | 0.75 | | | | 0.000 | 0.25 | | | || | | cmpq %rdx, %r12
|
||
|
|
1367 | 0.00 | | | | | | 1.00 | | | || | | setne %dil
|
||
|
|
1368 | 0.00 | 0.75 | | | | 0.000 | 0.25 | | | || | | leal 1(%rcx,%rcx), %ecx
|
||
|
|
1369 | 0.00 | | | | | | 1.00 | | | || | | sete %bpl
|
||
|
|
1370 | 0.00 | 0.50 | | | | 0.000 | 0.50 | | | || | | xorl %edx, %edx
|
||
|
|
1371 | 0.00 | 0.50 | | | | 0.000 | 0.50 | | | || | | xorl %ebx, %ebx
|
||
|
|
1372 | 0.00 | 0.75 | | | | 0.000 | 0.25 | | | || | | cmpq %rcx, %r12
|
||
|
|
1373 | 0.00 | | | | | | 1.00 | | | || | | sete %dl
|
||
|
|
1374 | 0.00 | 1.00 | | | | 0.000 | 0.00 | | | || | | movl $0, %ecx
|
||
|
|
1375 | 0.00 | | | | | | 1.00 | | | || | | setne %bl
|
||
|
|
1376 | 0.00 | | | | | | 1.00 | | | || | | cmovel %r8d, %ecx
|
||
|
|
1377 | 0.00 | 0.75 | | | | 0.000 | 0.25 | | | || | | movl %ebx, %r14d
|
||
|
|
1378 | 0.00 | | | | | | 1.00 | | | || | | shll $4, %r14d
|
||
|
|
1379 | 0.00 | 0.50 | | | | 0.000 | 0.50 | | | || | | subl %ebp, %r14d
|
||
|
|
1380 | 0.00 | 0.75 | | | | 0.000 | 0.25 | | | || | | leal (%rcx,%rdi,2), %ecx
|
||
|
|
1381 | 0.00 | | | | | | 1.00 | | | || | | shll $8, %ecx
|
||
|
|
1382 | 0.00 | 0.50 | | | | 0.000 | 0.50 | | | || | | addl $239, %r14d
|
||
|
|
1383 | 0.00 | 0.50 | | | | 0.000 | 0.50 | | | || | | addl $-768, %ecx # imm = 0xFD00
|
||
|
|
1384 | 0.00 | 1.00 | | | | 0.000 | 0.00 | | | || | | orl %r14d, %ecx
|
||
|
|
1385 | 1.00 | | | | | | | | | || | | kmovd %ecx, %k2
|
||
|
|
1386 | 0.50 | | | | | 0.500 | | | | || | | vcmpltps %zmm0, %zmm22, %k2 {%k2}
|
||
|
|
1387 | 0.50 | | | | | 0.500 | | | | || | | vsubps %zmm21, %zmm11, %zmm21
|
||
|
|
1388 | 0.50 | | | | | 0.500 | | | | || | | vsubps %zmm20, %zmm15, %zmm20
|
||
|
|
1389 | 0.50 | | | | | 0.500 | | | | || | | vsubps %zmm19, %zmm7, %zmm19
|
||
|
|
1390 | 0.50 | | | | | 0.500 | | | | || 4.0 | | vmulps %zmm2, %zmm23, %zmm22
|
||
|
|
1391 | 0.50 | | | | | 0.500 | | | | || 4.0 | | vfmadd231ps %zmm22, %zmm18, %zmm12 {%k2} # zmm12 {%k2} = (zmm18 * zmm22) + zmm12
|
||
|
|
1392 | 0.50 | | | | | 0.500 | | | | || | | vmulps %zmm20, %zmm20, %zmm18
|
||
|
|
1393 | 0.50 | | | | | 0.500 | | | | || | | vfmadd231ps %zmm21, %zmm21, %zmm18 # zmm18 = (zmm21 * zmm21) + zmm18
|
||
|
|
1394 | 0.50 | | | | | 0.500 | | | | || | | vfmadd231ps %zmm19, %zmm19, %zmm18 # zmm18 = (zmm19 * zmm19) + zmm18
|
||
|
|
1395 | 0.50 | | | | | 0.500 | | | | || | | vfmadd231ps %zmm22, %zmm17, %zmm9 {%k2} # zmm9 {%k2} = (zmm17 * zmm22) + zmm9
|
||
|
|
1396 | 2.50 | | | | | 0.500 | | | | || | | vrcp14ps %zmm18, %zmm17
|
||
|
|
1397 | 0.50 | | | | | 0.500 | | | | || | | vfmadd231ps %zmm22, %zmm16, %zmm5 {%k2} # zmm5 {%k2} = (zmm16 * zmm22) + zmm5
|
||
|
|
1398 | 0.50 | | | | | 0.500 | | | | || | | vmulps %zmm17, %zmm26, %zmm16
|
||
|
|
1399 | 0.50 | | | | | 0.500 | | | | || | | vmulps %zmm16, %zmm17, %zmm16
|
||
|
|
1400 | 0.50 | | | | | 0.500 | | | | || | | vmulps %zmm16, %zmm17, %zmm16
|
||
|
|
1401 | 0.00 | | | | | 1.000 | | | | || | | vaddps %zmm1, %zmm16, %zmm22
|
||
|
|
1402 | 0.00 | | | | | 1.000 | | | | || | | vmulps %zmm17, %zmm27, %zmm17
|
||
|
|
1403 | 0.00 | | | | | 1.000 | | | | || | | vmulps %zmm22, %zmm17, %zmm17
|
||
|
|
1404 | 0.00 | | | | | 1.000 | | | | || | | vmulps %zmm17, %zmm16, %zmm16
|
||
|
|
1405 | 0.00 | | | | | | 1.00 | | | || | | shll $6, %ebx
|
||
|
|
1406 | 0.00 | 1.00 | | | | 0.000 | 0.00 | | | || | | leal (%rbx,%rdi,4), %ecx
|
||
|
|
1407 | 0.00 | | | | | | 1.00 | | | || | | shll $7, %edx
|
||
|
|
1408 | 0.00 | 1.00 | | | | 0.000 | 0.00 | | | || | | leal (%rdx,%rdi,8), %edx
|
||
|
|
1409 | 0.00 | | | | | | 1.00 | | | || | | shll $8, %edx
|
||
|
|
1410 | 0.00 | 1.00 | | | | 0.000 | 0.00 | | | || | | addl %edx, %ecx
|
||
|
|
1411 | 0.00 | 1.00 | | | | 0.000 | 0.00 | | | || | | addl $-2117, %ecx # imm = 0xF7BB
|
||
|
|
1412 | 1.00 | | | | | | | | | || | | kmovd %ecx, %k2
|
||
|
|
1413 | 0.00 | | | | | 1.000 | | | | || | | vcmpltps %zmm0, %zmm18, %k2 {%k2}
|
||
|
|
1414 | 0.00 | | | | | 1.000 | | | | || | | vmulps %zmm2, %zmm16, %zmm16
|
||
|
|
1415 | 0.00 | | | | | 1.000 | | | | || | | vfmadd231ps %zmm16, %zmm19, %zmm13 {%k2} # zmm13 {%k2} = (zmm19 * zmm16) + zmm13
|
||
|
|
1416 | 0.00 | | | | | 1.000 | | | | || | | vfmadd231ps %zmm16, %zmm21, %zmm8 {%k2} # zmm8 {%k2} = (zmm21 * zmm16) + zmm8
|
||
|
|
1417 | 0.24 | | | | | 0.760 | | | | || | 4.0 | vfmadd231ps %zmm16, %zmm20, %zmm4 {%k2} # zmm4 {%k2} = (zmm20 * zmm16) + zmm4
|
||
|
|
1418 | 0.00 | 1.00 | | | | -0.01 | 0.00 | | | || | | incq %rax
|
||
|
|
1419 | 0.00 | 1.00 | | | | -0.01 | 0.00 | | | || | | cmpq %rax, %r10
|
||
|
|
1420 | | | | | | | | | | || | | * jne .LBB2_12
|
||
|
|
1421 | | | | | | | | | | || | | # LLVM-MCA-END
|
||
|
|
|
||
|
|
22.5 16.5 2.00 2.00 2.00 2.00 22.49 16.5 71 4.0
|
||
|
|
|
||
|
|
|
||
|
|
|
||
|
|
|
||
|
|
Loop-Carried Dependencies Analysis Report
|
||
|
|
-----------------------------------------
|
||
|
|
1417 | 4.0 | vfmadd231ps %zmm16, %zmm20, %zmm4 {%k2} # zmm4 {%k2} = (zmm20 * zmm16) + zmm4| [1417]
|
||
|
|
1416 | 4.0 | vfmadd231ps %zmm16, %zmm21, %zmm8 {%k2} # zmm8 {%k2} = (zmm21 * zmm16) + zmm8| [1416]
|
||
|
|
1415 | 4.0 | vfmadd231ps %zmm16, %zmm19, %zmm13 {%k2} # zmm13 {%k2} = (zmm19 * zmm16) + zmm13| [1415]
|
||
|
|
1397 | 4.0 | vfmadd231ps %zmm22, %zmm16, %zmm5 {%k2} # zmm5 {%k2} = (zmm16 * zmm22) + zmm5| [1397]
|
||
|
|
1395 | 4.0 | vfmadd231ps %zmm22, %zmm17, %zmm9 {%k2} # zmm9 {%k2} = (zmm17 * zmm22) + zmm9| [1395]
|
||
|
|
1391 | 4.0 | vfmadd231ps %zmm22, %zmm18, %zmm12 {%k2} # zmm12 {%k2} = (zmm18 * zmm22) + zmm12| [1391]
|
||
|
|
1418 | 1.0 | incq %rax | [1418]
|
||
|
|
|