[ARM] Adjust how NEON shifts are lowered
[llvm-core.git] / test / tools / llvm-mca / X86 / BdVer2 / dot-product.s
blob98cda18047050b3fa101012e19fda171e71ed482
1 # NOTE: Assertions have been autogenerated by utils/update_mca_test_checks.py
2 # RUN: llvm-mca -mtriple=x86_64-unknown-unknown -mcpu=bdver2 -iterations=300 -timeline -timeline-max-iterations=3 < %s | FileCheck %s
4 vmulps %xmm0, %xmm1, %xmm2
5 vhaddps %xmm2, %xmm2, %xmm3
6 vhaddps %xmm3, %xmm3, %xmm4
8 # CHECK: Iterations: 300
9 # CHECK-NEXT: Instructions: 900
10 # CHECK-NEXT: Total Cycles: 1709
11 # CHECK-NEXT: Total uOps: 2100
13 # CHECK: Dispatch Width: 4
14 # CHECK-NEXT: uOps Per Cycle: 1.23
15 # CHECK-NEXT: IPC: 0.53
16 # CHECK-NEXT: Block RThroughput: 5.5
18 # CHECK: Instruction Info:
19 # CHECK-NEXT: [1]: #uOps
20 # CHECK-NEXT: [2]: Latency
21 # CHECK-NEXT: [3]: RThroughput
22 # CHECK-NEXT: [4]: MayLoad
23 # CHECK-NEXT: [5]: MayStore
24 # CHECK-NEXT: [6]: HasSideEffects (U)
26 # CHECK: [1] [2] [3] [4] [5] [6] Instructions:
27 # CHECK-NEXT: 1 5 1.00 vmulps %xmm0, %xmm1, %xmm2
28 # CHECK-NEXT: 3 11 2.50 vhaddps %xmm2, %xmm2, %xmm3
29 # CHECK-NEXT: 3 11 2.50 vhaddps %xmm3, %xmm3, %xmm4
31 # CHECK: Resources:
32 # CHECK-NEXT: [0.0] - PdAGLU01
33 # CHECK-NEXT: [0.1] - PdAGLU01
34 # CHECK-NEXT: [1] - PdBranch
35 # CHECK-NEXT: [2] - PdCount
36 # CHECK-NEXT: [3] - PdDiv
37 # CHECK-NEXT: [4] - PdEX0
38 # CHECK-NEXT: [5] - PdEX1
39 # CHECK-NEXT: [6] - PdFPCVT
40 # CHECK-NEXT: [7.0] - PdFPFMA
41 # CHECK-NEXT: [7.1] - PdFPFMA
42 # CHECK-NEXT: [8.0] - PdFPMAL
43 # CHECK-NEXT: [8.1] - PdFPMAL
44 # CHECK-NEXT: [9] - PdFPMMA
45 # CHECK-NEXT: [10] - PdFPSTO
46 # CHECK-NEXT: [11] - PdFPU0
47 # CHECK-NEXT: [12] - PdFPU1
48 # CHECK-NEXT: [13] - PdFPU2
49 # CHECK-NEXT: [14] - PdFPU3
50 # CHECK-NEXT: [15] - PdFPXBR
51 # CHECK-NEXT: [16.0] - PdLoad
52 # CHECK-NEXT: [16.1] - PdLoad
53 # CHECK-NEXT: [17] - PdMul
54 # CHECK-NEXT: [18] - PdStore
56 # CHECK: Resource pressure per iteration:
57 # 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]
58 # CHECK-NEXT: - - - - - - - - 5.50 5.50 - - - - 2.00 1.00 - - - - - - -
60 # CHECK: Resource pressure by instruction:
61 # 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:
62 # CHECK-NEXT: - - - - - - - - 0.49 0.51 - - - - - 1.00 - - - - - - - vmulps %xmm0, %xmm1, %xmm2
63 # CHECK-NEXT: - - - - - - - - 2.53 2.47 - - - - 1.00 - - - - - - - - vhaddps %xmm2, %xmm2, %xmm3
64 # CHECK-NEXT: - - - - - - - - 2.48 2.52 - - - - 1.00 - - - - - - - - vhaddps %xmm3, %xmm3, %xmm4
66 # CHECK: Timeline view:
67 # CHECK-NEXT: 0123456789 01234
68 # CHECK-NEXT: Index 0123456789 0123456789
70 # CHECK: [0,0] DeeeeeER . . . . . . vmulps %xmm0, %xmm1, %xmm2
71 # CHECK-NEXT: [0,1] D=====eeeeeeeeeeeER . . . . vhaddps %xmm2, %xmm2, %xmm3
72 # CHECK-NEXT: [0,2] .D===============eeeeeeeeeeeER. . vhaddps %xmm3, %xmm3, %xmm4
73 # CHECK-NEXT: [1,0] .DeeeeeE---------------------R. . vmulps %xmm0, %xmm1, %xmm2
74 # CHECK-NEXT: [1,1] . D====eeeeeeeeeeeE----------R. . vhaddps %xmm2, %xmm2, %xmm3
75 # CHECK-NEXT: [1,2] . D==============eeeeeeeeeeeER . vhaddps %xmm3, %xmm3, %xmm4
76 # CHECK-NEXT: [2,0] . DeeeeeE--------------------R . vmulps %xmm0, %xmm1, %xmm2
77 # CHECK-NEXT: [2,1] . D======eeeeeeeeeeeE-------R . vhaddps %xmm2, %xmm2, %xmm3
78 # CHECK-NEXT: [2,2] . D================eeeeeeeeeeeER vhaddps %xmm3, %xmm3, %xmm4
80 # CHECK: Average Wait times (based on the timeline view):
81 # CHECK-NEXT: [0]: Executions
82 # CHECK-NEXT: [1]: Average time spent waiting in a scheduler's queue
83 # CHECK-NEXT: [2]: Average time spent waiting in a scheduler's queue while ready
84 # CHECK-NEXT: [3]: Average time elapsed from WB until retire stage
86 # CHECK: [0] [1] [2] [3]
87 # CHECK-NEXT: 0. 3 1.0 1.0 13.7 vmulps %xmm0, %xmm1, %xmm2
88 # CHECK-NEXT: 1. 3 6.0 0.7 5.7 vhaddps %xmm2, %xmm2, %xmm3
89 # CHECK-NEXT: 2. 3 16.0 0.0 0.0 vhaddps %xmm3, %xmm3, %xmm4