88 lines · plain
1# NOTE: Assertions have been autogenerated by utils/update_mca_test_checks.py2# RUN: llvm-mca -mtriple=x86_64-unknown-unknown -mcpu=znver1 -iterations=1500 -timeline -timeline-max-iterations=4 < %s | FileCheck %s3 4# Each lzcnt has a false dependency on %ecx; the first lzcnt has to wait on the5# imul. However, the folded load can start immediately.6# The last lzcnt has a false dependency on %cx. However, even in this case, the7# folded load can start immediately.8 9imul %edx, %ecx10lzcnt (%rsp), %cx11lzcnt 2(%rsp), %cx12 13# CHECK: Iterations: 150014# CHECK-NEXT: Instructions: 450015# CHECK-NEXT: Total Cycles: 1050316# CHECK-NEXT: Total uOps: 750017 18# CHECK: Dispatch Width: 419# CHECK-NEXT: uOps Per Cycle: 0.7120# CHECK-NEXT: IPC: 0.4321# CHECK-NEXT: Block RThroughput: 1.322 23# CHECK: Instruction Info:24# CHECK-NEXT: [1]: #uOps25# CHECK-NEXT: [2]: Latency26# CHECK-NEXT: [3]: RThroughput27# CHECK-NEXT: [4]: MayLoad28# CHECK-NEXT: [5]: MayStore29# CHECK-NEXT: [6]: HasSideEffects (U)30 31# CHECK: [1] [2] [3] [4] [5] [6] Instructions:32# CHECK-NEXT: 1 3 1.00 imull %edx, %ecx33# CHECK-NEXT: 2 6 0.50 * lzcntw (%rsp), %cx34# CHECK-NEXT: 2 6 0.50 * lzcntw 2(%rsp), %cx35 36# CHECK: Resources:37# CHECK-NEXT: [0] - ZnAGU038# CHECK-NEXT: [1] - ZnAGU139# CHECK-NEXT: [2] - ZnALU040# CHECK-NEXT: [3] - ZnALU141# CHECK-NEXT: [4] - ZnALU242# CHECK-NEXT: [5] - ZnALU343# CHECK-NEXT: [6] - ZnDivider44# CHECK-NEXT: [7] - ZnFPU045# CHECK-NEXT: [8] - ZnFPU146# CHECK-NEXT: [9] - ZnFPU247# CHECK-NEXT: [10] - ZnFPU348# CHECK-NEXT: [11] - ZnMultiplier49 50# CHECK: Resource pressure per iteration:51# CHECK-NEXT: [0] [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11]52# CHECK-NEXT: 1.00 1.00 0.67 1.00 0.67 0.67 - - - - - 1.0053 54# CHECK: Resource pressure by instruction:55# CHECK-NEXT: [0] [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] Instructions:56# CHECK-NEXT: - - - 1.00 - - - - - - - 1.00 imull %edx, %ecx57# CHECK-NEXT: - 1.00 0.33 - 0.33 0.33 - - - - - - lzcntw (%rsp), %cx58# CHECK-NEXT: 1.00 - 0.33 - 0.33 0.33 - - - - - - lzcntw 2(%rsp), %cx59 60# CHECK: Timeline view:61# CHECK-NEXT: 0123456789 062# CHECK-NEXT: Index 0123456789 012345678963 64# CHECK: [0,0] DeeeER . . . . . imull %edx, %ecx65# CHECK-NEXT: [0,1] DeeeeeeER . . . . . lzcntw (%rsp), %cx66# CHECK-NEXT: [0,2] .DeeeeeeER. . . . . lzcntw 2(%rsp), %cx67# CHECK-NEXT: [1,0] .D======eeeER . . . . imull %edx, %ecx68# CHECK-NEXT: [1,1] . D=====eeeeeeER . . . lzcntw (%rsp), %cx69# CHECK-NEXT: [1,2] . D======eeeeeeER . . . lzcntw 2(%rsp), %cx70# CHECK-NEXT: [2,0] . D===========eeeER. . . imull %edx, %ecx71# CHECK-NEXT: [2,1] . D===========eeeeeeER . . lzcntw (%rsp), %cx72# CHECK-NEXT: [2,2] . D===========eeeeeeER . . lzcntw 2(%rsp), %cx73# CHECK-NEXT: [3,0] . D=================eeeER . imull %edx, %ecx74# CHECK-NEXT: [3,1] . D================eeeeeeER. lzcntw (%rsp), %cx75# CHECK-NEXT: [3,2] . D=================eeeeeeER lzcntw 2(%rsp), %cx76 77# CHECK: Average Wait times (based on the timeline view):78# CHECK-NEXT: [0]: Executions79# CHECK-NEXT: [1]: Average time spent waiting in a scheduler's queue80# CHECK-NEXT: [2]: Average time spent waiting in a scheduler's queue while ready81# CHECK-NEXT: [3]: Average time elapsed from WB until retire stage82 83# CHECK: [0] [1] [2] [3]84# CHECK-NEXT: 0. 4 9.5 0.3 0.0 imull %edx, %ecx85# CHECK-NEXT: 1. 4 9.0 0.0 0.0 lzcntw (%rsp), %cx86# CHECK-NEXT: 2. 4 9.5 0.0 0.0 lzcntw 2(%rsp), %cx87# CHECK-NEXT: 4 9.3 0.1 0.0 <total>88