169 lines · plain
1# NOTE: Assertions have been autogenerated by utils/update_mca_test_checks.py2# RUN: llvm-mca -mtriple=x86_64-unknown-unknown -mcpu=bdver2 -timeline -timeline-max-iterations=1 -register-file-stats < %s | FileCheck %s3 4# These are dependency-breaking one-idioms.5# Much like zero-idioms, but they produce ones, and do consume resources.6 7# perf stats reports a throughput of 2.00 IPC.8 9pcmpeqb %mm2, %mm210pcmpeqd %mm2, %mm211pcmpeqw %mm2, %mm212 13pcmpeqb %xmm2, %xmm214pcmpeqd %xmm2, %xmm215pcmpeqq %xmm2, %xmm216pcmpeqw %xmm2, %xmm217 18vpcmpeqb %xmm3, %xmm3, %xmm319vpcmpeqd %xmm3, %xmm3, %xmm320vpcmpeqq %xmm3, %xmm3, %xmm321vpcmpeqw %xmm3, %xmm3, %xmm322 23vpcmpeqb %xmm3, %xmm3, %xmm524vpcmpeqd %xmm3, %xmm3, %xmm525vpcmpeqq %xmm3, %xmm3, %xmm526vpcmpeqw %xmm3, %xmm3, %xmm527 28# FIXME: their handling is broken in llvm-mca.29 30# CHECK: Iterations: 10031# CHECK-NEXT: Instructions: 150032# CHECK-NEXT: Total Cycles: 75433# CHECK-NEXT: Total uOps: 150034 35# CHECK: Dispatch Width: 436# CHECK-NEXT: uOps Per Cycle: 1.9937# CHECK-NEXT: IPC: 1.9938# CHECK-NEXT: Block RThroughput: 7.539 40# CHECK: Instruction Info:41# CHECK-NEXT: [1]: #uOps42# CHECK-NEXT: [2]: Latency43# CHECK-NEXT: [3]: RThroughput44# CHECK-NEXT: [4]: MayLoad45# CHECK-NEXT: [5]: MayStore46# CHECK-NEXT: [6]: HasSideEffects (U)47 48# CHECK: [1] [2] [3] [4] [5] [6] Instructions:49# CHECK-NEXT: 1 2 0.50 pcmpeqb %mm2, %mm250# CHECK-NEXT: 1 2 0.50 pcmpeqd %mm2, %mm251# CHECK-NEXT: 1 2 0.50 pcmpeqw %mm2, %mm252# CHECK-NEXT: 1 2 0.50 pcmpeqb %xmm2, %xmm253# CHECK-NEXT: 1 2 0.50 pcmpeqd %xmm2, %xmm254# CHECK-NEXT: 1 2 0.50 pcmpeqq %xmm2, %xmm255# CHECK-NEXT: 1 2 0.50 pcmpeqw %xmm2, %xmm256# CHECK-NEXT: 1 2 0.50 vpcmpeqb %xmm3, %xmm3, %xmm357# CHECK-NEXT: 1 2 0.50 vpcmpeqd %xmm3, %xmm3, %xmm358# CHECK-NEXT: 1 2 0.50 vpcmpeqq %xmm3, %xmm3, %xmm359# CHECK-NEXT: 1 2 0.50 vpcmpeqw %xmm3, %xmm3, %xmm360# CHECK-NEXT: 1 2 0.50 vpcmpeqb %xmm3, %xmm3, %xmm561# CHECK-NEXT: 1 2 0.50 vpcmpeqd %xmm3, %xmm3, %xmm562# CHECK-NEXT: 1 2 0.50 vpcmpeqq %xmm3, %xmm3, %xmm563# CHECK-NEXT: 1 2 0.50 vpcmpeqw %xmm3, %xmm3, %xmm564 65# CHECK: Register File statistics:66# CHECK-NEXT: Total number of mappings created: 150067# CHECK-NEXT: Max number of mappings used: 7268 69# CHECK: * Register File #1 -- PdFpuPRF:70# CHECK-NEXT: Number of physical registers: 16071# CHECK-NEXT: Total number of mappings created: 150072# CHECK-NEXT: Max number of mappings used: 7273 74# CHECK: * Register File #2 -- PdIntegerPRF:75# CHECK-NEXT: Number of physical registers: 9676# CHECK-NEXT: Total number of mappings created: 077# CHECK-NEXT: Max number of mappings used: 078 79# CHECK: Resources:80# CHECK-NEXT: [0.0] - PdAGLU0181# CHECK-NEXT: [0.1] - PdAGLU0182# CHECK-NEXT: [1] - PdBranch83# CHECK-NEXT: [2] - PdCount84# CHECK-NEXT: [3] - PdDiv85# CHECK-NEXT: [4] - PdEX086# CHECK-NEXT: [5] - PdEX187# CHECK-NEXT: [6] - PdFPCVT88# CHECK-NEXT: [7.0] - PdFPFMA89# CHECK-NEXT: [7.1] - PdFPFMA90# CHECK-NEXT: [8.0] - PdFPMAL91# CHECK-NEXT: [8.1] - PdFPMAL92# CHECK-NEXT: [9] - PdFPMMA93# CHECK-NEXT: [10] - PdFPSTO94# CHECK-NEXT: [11] - PdFPU095# CHECK-NEXT: [12] - PdFPU196# CHECK-NEXT: [13] - PdFPU297# CHECK-NEXT: [14] - PdFPU398# CHECK-NEXT: [15] - PdFPXBR99# CHECK-NEXT: [16.0] - PdLoad100# CHECK-NEXT: [16.1] - PdLoad101# CHECK-NEXT: [17] - PdMul102# CHECK-NEXT: [18] - PdStore103 104# CHECK: Resource pressure per iteration:105# CHECK-NEXT: [0.0] [0.1] [1] [2] [3] [4] [5] [6] [7.0] [7.1] [8.0] [8.1] [9] [10] [11] [12] [13] [14] [15] [16.0] [16.1] [17] [18]106# CHECK-NEXT: - - - - - - - - - - 7.50 7.50 - - - - 7.50 7.50 - - - - -107 108# CHECK: Resource pressure by instruction:109# CHECK-NEXT: [0.0] [0.1] [1] [2] [3] [4] [5] [6] [7.0] [7.1] [8.0] [8.1] [9] [10] [11] [12] [13] [14] [15] [16.0] [16.1] [17] [18] Instructions:110# CHECK-NEXT: - - - - - - - - - - 0.50 0.50 - - - - 0.50 0.50 - - - - - pcmpeqb %mm2, %mm2111# CHECK-NEXT: - - - - - - - - - - 0.50 0.50 - - - - 0.50 0.50 - - - - - pcmpeqd %mm2, %mm2112# CHECK-NEXT: - - - - - - - - - - 0.50 0.50 - - - - 0.50 0.50 - - - - - pcmpeqw %mm2, %mm2113# CHECK-NEXT: - - - - - - - - - - 0.50 0.50 - - - - 0.50 0.50 - - - - - pcmpeqb %xmm2, %xmm2114# CHECK-NEXT: - - - - - - - - - - 0.50 0.50 - - - - 0.50 0.50 - - - - - pcmpeqd %xmm2, %xmm2115# CHECK-NEXT: - - - - - - - - - - - 1.00 - - - - - 1.00 - - - - - pcmpeqq %xmm2, %xmm2116# CHECK-NEXT: - - - - - - - - - - 0.50 0.50 - - - - 0.50 0.50 - - - - - pcmpeqw %xmm2, %xmm2117# CHECK-NEXT: - - - - - - - - - - 0.50 0.50 - - - - 0.50 0.50 - - - - - vpcmpeqb %xmm3, %xmm3, %xmm3118# CHECK-NEXT: - - - - - - - - - - 1.00 - - - - - 1.00 - - - - - - vpcmpeqd %xmm3, %xmm3, %xmm3119# CHECK-NEXT: - - - - - - - - - - 0.50 0.50 - - - - 0.50 0.50 - - - - - vpcmpeqq %xmm3, %xmm3, %xmm3120# CHECK-NEXT: - - - - - - - - - - 0.50 0.50 - - - - 0.50 0.50 - - - - - vpcmpeqw %xmm3, %xmm3, %xmm3121# CHECK-NEXT: - - - - - - - - - - 0.50 0.50 - - - - 0.50 0.50 - - - - - vpcmpeqb %xmm3, %xmm3, %xmm5122# CHECK-NEXT: - - - - - - - - - - 0.50 0.50 - - - - 0.50 0.50 - - - - - vpcmpeqd %xmm3, %xmm3, %xmm5123# CHECK-NEXT: - - - - - - - - - - 0.50 0.50 - - - - 0.50 0.50 - - - - - vpcmpeqq %xmm3, %xmm3, %xmm5124# CHECK-NEXT: - - - - - - - - - - 0.50 0.50 - - - - 0.50 0.50 - - - - - vpcmpeqw %xmm3, %xmm3, %xmm5125 126# CHECK: Timeline view:127# CHECK-NEXT: 01128# CHECK-NEXT: Index 0123456789129 130# CHECK: [0,0] DeeER. .. pcmpeqb %mm2, %mm2131# CHECK-NEXT: [0,1] DeeER. .. pcmpeqd %mm2, %mm2132# CHECK-NEXT: [0,2] D=eeER .. pcmpeqw %mm2, %mm2133# CHECK-NEXT: [0,3] D==eeER .. pcmpeqb %xmm2, %xmm2134# CHECK-NEXT: [0,4] .DeeE-R .. pcmpeqd %xmm2, %xmm2135# CHECK-NEXT: [0,5] .D==eeER .. pcmpeqq %xmm2, %xmm2136# CHECK-NEXT: [0,6] .D=eeE-R .. pcmpeqw %xmm2, %xmm2137# CHECK-NEXT: [0,7] .D===eeER .. vpcmpeqb %xmm3, %xmm3, %xmm3138# CHECK-NEXT: [0,8] . D=eeE-R .. vpcmpeqd %xmm3, %xmm3, %xmm3139# CHECK-NEXT: [0,9] . D===eeER.. vpcmpeqq %xmm3, %xmm3, %xmm3140# CHECK-NEXT: [0,10] . D==eeE-R.. vpcmpeqw %xmm3, %xmm3, %xmm3141# CHECK-NEXT: [0,11] . D===eeER.. vpcmpeqb %xmm3, %xmm3, %xmm5142# CHECK-NEXT: [0,12] . D===eeER. vpcmpeqd %xmm3, %xmm3, %xmm5143# CHECK-NEXT: [0,13] . D===eeER. vpcmpeqq %xmm3, %xmm3, %xmm5144# CHECK-NEXT: [0,14] . D====eeER vpcmpeqw %xmm3, %xmm3, %xmm5145 146# CHECK: Average Wait times (based on the timeline view):147# CHECK-NEXT: [0]: Executions148# CHECK-NEXT: [1]: Average time spent waiting in a scheduler's queue149# CHECK-NEXT: [2]: Average time spent waiting in a scheduler's queue while ready150# CHECK-NEXT: [3]: Average time elapsed from WB until retire stage151 152# CHECK: [0] [1] [2] [3]153# CHECK-NEXT: 0. 1 1.0 1.0 0.0 pcmpeqb %mm2, %mm2154# CHECK-NEXT: 1. 1 1.0 1.0 0.0 pcmpeqd %mm2, %mm2155# CHECK-NEXT: 2. 1 2.0 2.0 0.0 pcmpeqw %mm2, %mm2156# CHECK-NEXT: 3. 1 3.0 3.0 0.0 pcmpeqb %xmm2, %xmm2157# CHECK-NEXT: 4. 1 1.0 1.0 1.0 pcmpeqd %xmm2, %xmm2158# CHECK-NEXT: 5. 1 3.0 0.0 0.0 pcmpeqq %xmm2, %xmm2159# CHECK-NEXT: 6. 1 2.0 2.0 1.0 pcmpeqw %xmm2, %xmm2160# CHECK-NEXT: 7. 1 4.0 4.0 0.0 vpcmpeqb %xmm3, %xmm3, %xmm3161# CHECK-NEXT: 8. 1 2.0 2.0 1.0 vpcmpeqd %xmm3, %xmm3, %xmm3162# CHECK-NEXT: 9. 1 4.0 0.0 0.0 vpcmpeqq %xmm3, %xmm3, %xmm3163# CHECK-NEXT: 10. 1 3.0 3.0 1.0 vpcmpeqw %xmm3, %xmm3, %xmm3164# CHECK-NEXT: 11. 1 4.0 4.0 0.0 vpcmpeqb %xmm3, %xmm3, %xmm5165# CHECK-NEXT: 12. 1 4.0 4.0 0.0 vpcmpeqd %xmm3, %xmm3, %xmm5166# CHECK-NEXT: 13. 1 4.0 0.0 0.0 vpcmpeqq %xmm3, %xmm3, %xmm5167# CHECK-NEXT: 14. 1 5.0 5.0 0.0 vpcmpeqw %xmm3, %xmm3, %xmm5168# CHECK-NEXT: 1 2.9 2.1 0.3 <total>169