test/tools/llvm-mca/X86/bextr-read-after-ld.s

   1 # NOTE: Assertions have been autogenerated by utils/update_mca_test_checks.py
   2 # RUN: llvm-mca -mtriple=x86_64-unknown-unknown -mcpu=haswell -iterations=1 -timeline -resource-pressure=false < %s | FileCheck %s -check-prefix=ALL -check-prefix=HASWELL
   3 # RUN: llvm-mca -mtriple=x86_64-unknown-unknown -mcpu=broadwell -iterations=1 -timeline -resource-pressure=false < %s | FileCheck %s -check-prefix=ALL -check-prefix=BDWELL
   4 # RUN: llvm-mca -mtriple=x86_64-unknown-unknown -mcpu=skylake -iterations=1 -timeline -resource-pressure=false < %s | FileCheck %s -check-prefix=ALL -check-prefix=SKYLAKE
   5 # RUN: llvm-mca -mtriple=x86_64-unknown-unknown -mcpu=bdver2 -iterations=1 -timeline -resource-pressure=false < %s | FileCheck %s -check-prefix=ALL -check-prefix=BDVER2
   6 # RUN: llvm-mca -mtriple=x86_64-unknown-unknown -mcpu=btver2 -iterations=1 -timeline -resource-pressure=false < %s | FileCheck %s -check-prefix=ALL -check-prefix=BTVER2
   7 # RUN: llvm-mca -mtriple=x86_64-unknown-unknown -mcpu=znver1 -iterations=1 -timeline -resource-pressure=false < %s | FileCheck %s -check-prefix=ALL -check-prefix=ZNVER1
   8
   9 add     %edi, %esi
  10 bextrl  %esi, (%rdi), %eax
  11
  12 # ALL:          Iterations:        1
  13 # ALL-NEXT:     Instructions:      2
  14
  15 # BDVER2-NEXT:  Total Cycles:      9
  16 # BDVER2-NEXT:  Total uOps:        3
  17
  18 # BDWELL-NEXT:  Total Cycles:      10
  19 # BDWELL-NEXT:  Total uOps:        4
  20
  21 # BTVER2-NEXT:  Total Cycles:      7
  22 # BTVER2-NEXT:  Total uOps:        2
  23
  24 # HASWELL-NEXT: Total Cycles:      10
  25 # HASWELL-NEXT: Total uOps:        4
  26
  27 # SKYLAKE-NEXT: Total Cycles:      10
  28 # SKYLAKE-NEXT: Total uOps:        4
  29
  30 # ZNVER1-NEXT:  Total Cycles:      8
  31 # ZNVER1-NEXT:  Total uOps:        3
  32
  33 # BDVER2:       Dispatch Width:    4
  34 # BDVER2-NEXT:  uOps Per Cycle:    0.33
  35 # BDVER2-NEXT:  IPC:               0.22
  36 # BDVER2-NEXT:  Block RThroughput: 2.0
  37
  38 # BDWELL:       Dispatch Width:    4
  39 # BDWELL-NEXT:  uOps Per Cycle:    0.40
  40 # BDWELL-NEXT:  IPC:               0.20
  41 # BDWELL-NEXT:  Block RThroughput: 1.0
  42
  43 # BTVER2:       Dispatch Width:    2
  44 # BTVER2-NEXT:  uOps Per Cycle:    0.29
  45 # BTVER2-NEXT:  IPC:               0.29
  46 # BTVER2-NEXT:  Block RThroughput: 1.0
  47
  48 # HASWELL:      Dispatch Width:    4
  49 # HASWELL-NEXT: uOps Per Cycle:    0.40
  50 # HASWELL-NEXT: IPC:               0.20
  51 # HASWELL-NEXT: Block RThroughput: 1.0
  52
  53 # SKYLAKE:      Dispatch Width:    6
  54 # SKYLAKE-NEXT: uOps Per Cycle:    0.40
  55 # SKYLAKE-NEXT: IPC:               0.20
  56 # SKYLAKE-NEXT: Block RThroughput: 0.7
  57
  58 # ZNVER1:       Dispatch Width:    4
  59 # ZNVER1-NEXT:  uOps Per Cycle:    0.38
  60 # ZNVER1-NEXT:  IPC:               0.25
  61 # ZNVER1-NEXT:  Block RThroughput: 0.8
  62
  63 # ALL:          Instruction Info:
  64 # ALL-NEXT:     [1]: #uOps
  65 # ALL-NEXT:     [2]: Latency
  66 # ALL-NEXT:     [3]: RThroughput
  67 # ALL-NEXT:     [4]: MayLoad
  68 # ALL-NEXT:     [5]: MayStore
  69 # ALL-NEXT:     [6]: HasSideEffects (U)
  70
  71 # ALL:          [1]    [2]    [3]    [4]    [5]    [6]    Instructions:
  72
  73 # BDVER2-NEXT:   1      1     1.00                        addl  %edi, %esi
  74 # BDVER2-NEXT:   2      6     1.50    *                   bextrl        %esi, (%rdi), %eax
  75
  76 # BDWELL-NEXT:   1      1     0.25                        addl  %edi, %esi
  77 # BDWELL-NEXT:   3      7     0.50    *                   bextrl        %esi, (%rdi), %eax
  78
  79 # BTVER2-NEXT:   1      1     0.50                        addl  %edi, %esi
  80 # BTVER2-NEXT:   1      4     1.00    *                   bextrl        %esi, (%rdi), %eax
  81
  82 # HASWELL-NEXT:  1      1     0.25                        addl  %edi, %esi
  83 # HASWELL-NEXT:  3      7     0.50    *                   bextrl        %esi, (%rdi), %eax
  84
  85 # SKYLAKE-NEXT:  1      1     0.25                        addl  %edi, %esi
  86 # SKYLAKE-NEXT:  3      7     0.50    *                   bextrl        %esi, (%rdi), %eax
  87
  88 # ZNVER1-NEXT:   1      1     0.25                        addl  %edi, %esi
  89 # ZNVER1-NEXT:   2      5     0.50    *                   bextrl        %esi, (%rdi), %eax
  90
  91 # ALL:          Timeline view:
  92
  93 # BDVER2-NEXT:  Index     012345678
  94 # BDWELL-NEXT:  Index     0123456789
  95 # BTVER2-NEXT:  Index     0123456
  96 # HASWELL-NEXT: Index     0123456789
  97 # SKYLAKE-NEXT: Index     0123456789
  98 # ZNVER1-NEXT:  Index     01234567
  99
 100 # BDVER2:       [0,0]     DeER .  .   addl      %edi, %esi
 101 # BDVER2-NEXT:  [0,1]     DeeeeeeER   bextrl    %esi, (%rdi), %eax
 102
 103 # BDWELL:       [0,0]     DeER .   .   addl     %edi, %esi
 104 # BDWELL-NEXT:  [0,1]     DeeeeeeeER   bextrl   %esi, (%rdi), %eax
 105
 106 # BTVER2:       [0,0]     DeER ..   addl        %edi, %esi
 107 # BTVER2-NEXT:  [0,1]     DeeeeER   bextrl      %esi, (%rdi), %eax
 108
 109 # HASWELL:      [0,0]     DeER .   .   addl     %edi, %esi
 110 # HASWELL-NEXT: [0,1]     DeeeeeeeER   bextrl   %esi, (%rdi), %eax
 111
 112 # SKYLAKE:      [0,0]     DeER .   .   addl     %edi, %esi
 113 # SKYLAKE-NEXT: [0,1]     DeeeeeeeER   bextrl   %esi, (%rdi), %eax
 114
 115 # ZNVER1:       [0,0]     DeER . .   addl       %edi, %esi
 116 # ZNVER1-NEXT:  [0,1]     DeeeeeER   bextrl     %esi, (%rdi), %eax
 117
 118 # ALL:          Average Wait times (based on the timeline view):
 119 # ALL-NEXT:     [0]: Executions
 120 # ALL-NEXT:     [1]: Average time spent waiting in a scheduler's queue
 121 # ALL-NEXT:     [2]: Average time spent waiting in a scheduler's queue while ready
 122 # ALL-NEXT:     [3]: Average time elapsed from WB until retire stage
 123
 124 # ALL:                [0]    [1]    [2]    [3]
 125 # ALL-NEXT:     0.     1     1.0    1.0    0.0       addl       %edi, %esi
 126 # ALL-NEXT:     1.     1     1.0    0.0    0.0       bextrl     %esi, (%rdi), %eax