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=btver2 -iterations=1500 -timeline -timeline-max-iterations=3 < %s | FileCheck %s3 4# perf stat reports a throughput of 0.60 IPC for this code snippet.5# Each lzcnt has a false dependency on %ecx; the first lzcnt has to wait on the6# imul. However, the folded load can start immediately.7# The last lzcnt has a false dependency on %cx. However, even in this case, the8# folded load can start immediately.9 10imul %edx, %ecx11lzcnt (%rsp), %cx12lzcnt 2(%rsp), %cx13 14# CHECK: Iterations: 150015# CHECK-NEXT: Instructions: 450016# CHECK-NEXT: Total Cycles: 750317# CHECK-NEXT: Total uOps: 450018 19# CHECK: Dispatch Width: 220# CHECK-NEXT: uOps Per Cycle: 0.6021# CHECK-NEXT: IPC: 0.6022# CHECK-NEXT: Block RThroughput: 2.023 24# CHECK: Instruction Info:25# CHECK-NEXT: [1]: #uOps26# CHECK-NEXT: [2]: Latency27# CHECK-NEXT: [3]: RThroughput28# CHECK-NEXT: [4]: MayLoad29# CHECK-NEXT: [5]: MayStore30# CHECK-NEXT: [6]: HasSideEffects (U)31 32# CHECK: [1] [2] [3] [4] [5] [6] Instructions:33# CHECK-NEXT: 1 3 1.00 imull %edx, %ecx34# CHECK-NEXT: 1 4 1.00 * lzcntw (%rsp), %cx35# CHECK-NEXT: 1 4 1.00 * lzcntw 2(%rsp), %cx36 37# CHECK: Resources:38# CHECK-NEXT: [0] - JALU039# CHECK-NEXT: [1] - JALU140# CHECK-NEXT: [2] - JDiv41# CHECK-NEXT: [3] - JFPA42# CHECK-NEXT: [4] - JFPM43# CHECK-NEXT: [5] - JFPU044# CHECK-NEXT: [6] - JFPU145# CHECK-NEXT: [7] - JLAGU46# CHECK-NEXT: [8] - JMul47# CHECK-NEXT: [9] - JSAGU48# CHECK-NEXT: [10] - JSTC49# CHECK-NEXT: [11] - JVALU050# CHECK-NEXT: [12] - JVALU151# CHECK-NEXT: [13] - JVIMUL52 53# CHECK: Resource pressure per iteration:54# CHECK-NEXT: [0] [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13]55# CHECK-NEXT: 1.50 1.50 - - - - - 2.00 1.00 - - - - -56 57# CHECK: Resource pressure by instruction:58# CHECK-NEXT: [0] [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] Instructions:59# CHECK-NEXT: - 1.00 - - - - - - 1.00 - - - - - imull %edx, %ecx60# CHECK-NEXT: 1.00 - - - - - - 1.00 - - - - - - lzcntw (%rsp), %cx61# CHECK-NEXT: 0.50 0.50 - - - - - 1.00 - - - - - - lzcntw 2(%rsp), %cx62 63# CHECK: Timeline view:64# CHECK-NEXT: 0123456765# CHECK-NEXT: Index 012345678966 67# CHECK: [0,0] DeeeER . . . imull %edx, %ecx68# CHECK-NEXT: [0,1] DeeeeER . . . lzcntw (%rsp), %cx69# CHECK-NEXT: [0,2] .DeeeeER . . . lzcntw 2(%rsp), %cx70# CHECK-NEXT: [1,0] .D====eeeER . . imull %edx, %ecx71# CHECK-NEXT: [1,1] . D===eeeeER . . lzcntw (%rsp), %cx72# CHECK-NEXT: [1,2] . D====eeeeER . . lzcntw 2(%rsp), %cx73# CHECK-NEXT: [2,0] . D=======eeeER . imull %edx, %ecx74# CHECK-NEXT: [2,1] . D=======eeeeER. lzcntw (%rsp), %cx75# CHECK-NEXT: [2,2] . D=======eeeeER 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. 3 4.7 0.3 0.0 imull %edx, %ecx85# CHECK-NEXT: 1. 3 4.3 0.0 0.0 lzcntw (%rsp), %cx86# CHECK-NEXT: 2. 3 4.7 0.0 0.0 lzcntw 2(%rsp), %cx87# CHECK-NEXT: 3 4.6 0.1 0.0 <total>88