From cd3cced8aa36d4773f48920fd2c04043b8d7e010 Mon Sep 17 00:00:00 2001 From: Simon Pilgrim Date: Sat, 12 Nov 2022 17:38:29 +0000 Subject: [MCA][X86] Add test coverage for VNNI instructions --- .../llvm-mca/X86/AlderlakeP/resources-avxvnni.s | 90 ++++++++ .../llvm-mca/X86/Generic/resources-avx512vnni.s | 141 +++++++++++ .../llvm-mca/X86/Generic/resources-avx512vnnivl.s | 253 ++++++++++++++++++++ .../tools/llvm-mca/X86/Generic/resources-avxvnni.s | 85 +++++++ .../X86/IceLakeServer/resources-avx512vnni.s | 145 ++++++++++++ .../X86/IceLakeServer/resources-avx512vnnivl.s | 257 +++++++++++++++++++++ .../llvm-mca/X86/IceLakeServer/resources-avxvnni.s | 89 +++++++ 7 files changed, 1060 insertions(+) create mode 100644 llvm/test/tools/llvm-mca/X86/AlderlakeP/resources-avxvnni.s create mode 100644 llvm/test/tools/llvm-mca/X86/Generic/resources-avx512vnni.s create mode 100644 llvm/test/tools/llvm-mca/X86/Generic/resources-avx512vnnivl.s create mode 100644 llvm/test/tools/llvm-mca/X86/Generic/resources-avxvnni.s create mode 100644 llvm/test/tools/llvm-mca/X86/IceLakeServer/resources-avx512vnni.s create mode 100644 llvm/test/tools/llvm-mca/X86/IceLakeServer/resources-avx512vnnivl.s create mode 100644 llvm/test/tools/llvm-mca/X86/IceLakeServer/resources-avxvnni.s diff --git a/llvm/test/tools/llvm-mca/X86/AlderlakeP/resources-avxvnni.s b/llvm/test/tools/llvm-mca/X86/AlderlakeP/resources-avxvnni.s new file mode 100644 index 000000000000..5c8d5e74e7ed --- /dev/null +++ b/llvm/test/tools/llvm-mca/X86/AlderlakeP/resources-avxvnni.s @@ -0,0 +1,90 @@ +# NOTE: Assertions have been autogenerated by utils/update_mca_test_checks.py +# RUN: llvm-mca -mtriple=x86_64-unknown-unknown -mcpu=alderlake -instruction-tables < %s | FileCheck %s + +vpdpbusd %xmm0, %xmm1, %xmm2 +vpdpbusd (%rax), %xmm1, %xmm2 + +vpdpbusd %ymm0, %ymm1, %ymm2 +vpdpbusd (%rax), %ymm1, %ymm2 + +vpdpbusds %xmm0, %xmm1, %xmm2 +vpdpbusds (%rax), %xmm1, %xmm2 + +vpdpbusds %ymm0, %ymm1, %ymm2 +vpdpbusds (%rax), %ymm1, %ymm2 + +vpdpwssd %xmm0, %xmm1, %xmm2 +vpdpwssd (%rax), %xmm1, %xmm2 + +vpdpwssd %ymm0, %ymm1, %ymm2 +vpdpwssd (%rax), %ymm1, %ymm2 + +vpdpwssds %xmm0, %xmm1, %xmm2 +vpdpwssds (%rax), %xmm1, %xmm2 + +vpdpwssds %ymm0, %ymm1, %ymm2 +vpdpwssds (%rax), %ymm1, %ymm2 + +# CHECK: Instruction Info: +# CHECK-NEXT: [1]: #uOps +# CHECK-NEXT: [2]: Latency +# CHECK-NEXT: [3]: RThroughput +# CHECK-NEXT: [4]: MayLoad +# CHECK-NEXT: [5]: MayStore +# CHECK-NEXT: [6]: HasSideEffects (U) + +# CHECK: [1] [2] [3] [4] [5] [6] Instructions: +# CHECK-NEXT: 1 5 0.50 vpdpbusd %xmm0, %xmm1, %xmm2 +# CHECK-NEXT: 2 13 0.50 * vpdpbusd (%rax), %xmm1, %xmm2 +# CHECK-NEXT: 1 5 0.50 vpdpbusd %ymm0, %ymm1, %ymm2 +# CHECK-NEXT: 2 13 0.50 * vpdpbusd (%rax), %ymm1, %ymm2 +# CHECK-NEXT: 1 5 0.50 vpdpbusds %xmm0, %xmm1, %xmm2 +# CHECK-NEXT: 2 13 0.50 * vpdpbusds (%rax), %xmm1, %xmm2 +# CHECK-NEXT: 1 5 0.50 vpdpbusds %ymm0, %ymm1, %ymm2 +# CHECK-NEXT: 2 13 0.50 * vpdpbusds (%rax), %ymm1, %ymm2 +# CHECK-NEXT: 1 5 0.50 vpdpwssd %xmm0, %xmm1, %xmm2 +# CHECK-NEXT: 2 13 0.50 * vpdpwssd (%rax), %xmm1, %xmm2 +# CHECK-NEXT: 1 5 0.50 vpdpwssd %ymm0, %ymm1, %ymm2 +# CHECK-NEXT: 2 13 0.50 * vpdpwssd (%rax), %ymm1, %ymm2 +# CHECK-NEXT: 1 5 0.50 vpdpwssds %xmm0, %xmm1, %xmm2 +# CHECK-NEXT: 2 13 0.50 * vpdpwssds (%rax), %xmm1, %xmm2 +# CHECK-NEXT: 1 5 0.50 vpdpwssds %ymm0, %ymm1, %ymm2 +# CHECK-NEXT: 2 13 0.50 * vpdpwssds (%rax), %ymm1, %ymm2 + +# CHECK: Resources: +# CHECK-NEXT: [0] - ADLPPort00 +# CHECK-NEXT: [1] - ADLPPort01 +# CHECK-NEXT: [2] - ADLPPort02 +# CHECK-NEXT: [3] - ADLPPort03 +# CHECK-NEXT: [4] - ADLPPort04 +# CHECK-NEXT: [5] - ADLPPort05 +# CHECK-NEXT: [6] - ADLPPort06 +# CHECK-NEXT: [7] - ADLPPort07 +# CHECK-NEXT: [8] - ADLPPort08 +# CHECK-NEXT: [9] - ADLPPort09 +# CHECK-NEXT: [10] - ADLPPort10 +# CHECK-NEXT: [11] - ADLPPort11 +# CHECK-NEXT: [12] - ADLPPortInvalid + +# CHECK: Resource pressure per iteration: +# CHECK-NEXT: [0] [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] +# CHECK-NEXT: 8.00 8.00 2.67 2.67 - - - - - - - 2.67 - + +# CHECK: Resource pressure by instruction: +# CHECK-NEXT: [0] [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] Instructions: +# CHECK-NEXT: 0.50 0.50 - - - - - - - - - - - vpdpbusd %xmm0, %xmm1, %xmm2 +# CHECK-NEXT: 0.50 0.50 0.33 0.33 - - - - - - - 0.33 - vpdpbusd (%rax), %xmm1, %xmm2 +# CHECK-NEXT: 0.50 0.50 - - - - - - - - - - - vpdpbusd %ymm0, %ymm1, %ymm2 +# CHECK-NEXT: 0.50 0.50 0.33 0.33 - - - - - - - 0.33 - vpdpbusd (%rax), %ymm1, %ymm2 +# CHECK-NEXT: 0.50 0.50 - - - - - - - - - - - vpdpbusds %xmm0, %xmm1, %xmm2 +# CHECK-NEXT: 0.50 0.50 0.33 0.33 - - - - - - - 0.33 - vpdpbusds (%rax), %xmm1, %xmm2 +# CHECK-NEXT: 0.50 0.50 - - - - - - - - - - - vpdpbusds %ymm0, %ymm1, %ymm2 +# CHECK-NEXT: 0.50 0.50 0.33 0.33 - - - - - - - 0.33 - vpdpbusds (%rax), %ymm1, %ymm2 +# CHECK-NEXT: 0.50 0.50 - - - - - - - - - - - vpdpwssd %xmm0, %xmm1, %xmm2 +# CHECK-NEXT: 0.50 0.50 0.33 0.33 - - - - - - - 0.33 - vpdpwssd (%rax), %xmm1, %xmm2 +# CHECK-NEXT: 0.50 0.50 - - - - - - - - - - - vpdpwssd %ymm0, %ymm1, %ymm2 +# CHECK-NEXT: 0.50 0.50 0.33 0.33 - - - - - - - 0.33 - vpdpwssd (%rax), %ymm1, %ymm2 +# CHECK-NEXT: 0.50 0.50 - - - - - - - - - - - vpdpwssds %xmm0, %xmm1, %xmm2 +# CHECK-NEXT: 0.50 0.50 0.33 0.33 - - - - - - - 0.33 - vpdpwssds (%rax), %xmm1, %xmm2 +# CHECK-NEXT: 0.50 0.50 - - - - - - - - - - - vpdpwssds %ymm0, %ymm1, %ymm2 +# CHECK-NEXT: 0.50 0.50 0.33 0.33 - - - - - - - 0.33 - vpdpwssds (%rax), %ymm1, %ymm2 diff --git a/llvm/test/tools/llvm-mca/X86/Generic/resources-avx512vnni.s b/llvm/test/tools/llvm-mca/X86/Generic/resources-avx512vnni.s new file mode 100644 index 000000000000..f1d0d4f7064c --- /dev/null +++ b/llvm/test/tools/llvm-mca/X86/Generic/resources-avx512vnni.s @@ -0,0 +1,141 @@ +# NOTE: Assertions have been autogenerated by utils/update_mca_test_checks.py +# RUN: llvm-mca -mtriple=x86_64-unknown-unknown -mcpu=x86-64 -instruction-tables < %s | FileCheck %s + +vpdpbusd %zmm0, %zmm1, %zmm2 +vpdpbusd (%rax), %zmm1, %zmm2 +vpdpbusd (%rax){1to16}, %zmm1, %zmm2 +vpdpbusd %zmm0, %zmm1, %zmm2 {k1} +vpdpbusd (%rax), %zmm1, %zmm2 {k1} +vpdpbusd (%rax){1to16}, %zmm1, %zmm2 {k1} +vpdpbusd %zmm0, %zmm1, %zmm2 {z}{k1} +vpdpbusd (%rax), %zmm1, %zmm2 {z}{k1} +vpdpbusd (%rax){1to16}, %zmm1, %zmm2 {z}{k1} + +vpdpbusds %zmm0, %zmm1, %zmm2 +vpdpbusds (%rax), %zmm1, %zmm2 +vpdpbusds (%rax){1to16}, %zmm1, %zmm2 +vpdpbusds %zmm0, %zmm1, %zmm2 {k1} +vpdpbusds (%rax), %zmm1, %zmm2 {k1} +vpdpbusds (%rax){1to16}, %zmm1, %zmm2 {k1} +vpdpbusds %zmm0, %zmm1, %zmm2 {z}{k1} +vpdpbusds (%rax), %zmm1, %zmm2 {z}{k1} +vpdpbusds (%rax){1to16}, %zmm1, %zmm2 {z}{k1} + +vpdpwssd %zmm0, %zmm1, %zmm2 +vpdpwssd (%rax), %zmm1, %zmm2 +vpdpwssd (%rax){1to16}, %zmm1, %zmm2 +vpdpwssd %zmm0, %zmm1, %zmm2 {k1} +vpdpwssd (%rax), %zmm1, %zmm2 {k1} +vpdpwssd (%rax){1to16}, %zmm1, %zmm2 {k1} +vpdpwssd %zmm0, %zmm1, %zmm2 {z}{k1} +vpdpwssd (%rax), %zmm1, %zmm2 {z}{k1} +vpdpwssd (%rax){1to16}, %zmm1, %zmm2 {z}{k1} + +vpdpwssds %zmm0, %zmm1, %zmm2 +vpdpwssds (%rax), %zmm1, %zmm2 +vpdpwssds (%rax){1to16}, %zmm1, %zmm2 +vpdpwssds %zmm0, %zmm1, %zmm2 {k1} +vpdpwssds (%rax), %zmm1, %zmm2 {k1} +vpdpwssds (%rax){1to16}, %zmm1, %zmm2 {k1} +vpdpwssds %zmm0, %zmm1, %zmm2 {z}{k1} +vpdpwssds (%rax), %zmm1, %zmm2 {z}{k1} +vpdpwssds (%rax){1to16}, %zmm1, %zmm2 {z}{k1} + +# CHECK: Instruction Info: +# CHECK-NEXT: [1]: #uOps +# CHECK-NEXT: [2]: Latency +# CHECK-NEXT: [3]: RThroughput +# CHECK-NEXT: [4]: MayLoad +# CHECK-NEXT: [5]: MayStore +# CHECK-NEXT: [6]: HasSideEffects (U) + +# CHECK: [1] [2] [3] [4] [5] [6] Instructions: +# CHECK-NEXT: 1 5 1.00 vpdpbusd %zmm0, %zmm1, %zmm2 +# CHECK-NEXT: 2 12 1.00 * vpdpbusd (%rax), %zmm1, %zmm2 +# CHECK-NEXT: 2 12 1.00 * vpdpbusd (%rax){1to16}, %zmm1, %zmm2 +# CHECK-NEXT: 1 5 1.00 vpdpbusd %zmm0, %zmm1, %zmm2 {%k1} +# CHECK-NEXT: 2 12 1.00 * vpdpbusd (%rax), %zmm1, %zmm2 {%k1} +# CHECK-NEXT: 2 12 1.00 * vpdpbusd (%rax){1to16}, %zmm1, %zmm2 {%k1} +# CHECK-NEXT: 1 5 1.00 vpdpbusd %zmm0, %zmm1, %zmm2 {%k1} {z} +# CHECK-NEXT: 2 12 1.00 * vpdpbusd (%rax), %zmm1, %zmm2 {%k1} {z} +# CHECK-NEXT: 2 12 1.00 * vpdpbusd (%rax){1to16}, %zmm1, %zmm2 {%k1} {z} +# CHECK-NEXT: 1 5 1.00 vpdpbusds %zmm0, %zmm1, %zmm2 +# CHECK-NEXT: 2 12 1.00 * vpdpbusds (%rax), %zmm1, %zmm2 +# CHECK-NEXT: 2 12 1.00 * vpdpbusds (%rax){1to16}, %zmm1, %zmm2 +# CHECK-NEXT: 1 5 1.00 vpdpbusds %zmm0, %zmm1, %zmm2 {%k1} +# CHECK-NEXT: 2 12 1.00 * vpdpbusds (%rax), %zmm1, %zmm2 {%k1} +# CHECK-NEXT: 2 12 1.00 * vpdpbusds (%rax){1to16}, %zmm1, %zmm2 {%k1} +# CHECK-NEXT: 1 5 1.00 vpdpbusds %zmm0, %zmm1, %zmm2 {%k1} {z} +# CHECK-NEXT: 2 12 1.00 * vpdpbusds (%rax), %zmm1, %zmm2 {%k1} {z} +# CHECK-NEXT: 2 12 1.00 * vpdpbusds (%rax){1to16}, %zmm1, %zmm2 {%k1} {z} +# CHECK-NEXT: 1 5 1.00 vpdpwssd %zmm0, %zmm1, %zmm2 +# CHECK-NEXT: 2 12 1.00 * vpdpwssd (%rax), %zmm1, %zmm2 +# CHECK-NEXT: 2 12 1.00 * vpdpwssd (%rax){1to16}, %zmm1, %zmm2 +# CHECK-NEXT: 1 5 1.00 vpdpwssd %zmm0, %zmm1, %zmm2 {%k1} +# CHECK-NEXT: 2 12 1.00 * vpdpwssd (%rax), %zmm1, %zmm2 {%k1} +# CHECK-NEXT: 2 12 1.00 * vpdpwssd (%rax){1to16}, %zmm1, %zmm2 {%k1} +# CHECK-NEXT: 1 5 1.00 vpdpwssd %zmm0, %zmm1, %zmm2 {%k1} {z} +# CHECK-NEXT: 2 12 1.00 * vpdpwssd (%rax), %zmm1, %zmm2 {%k1} {z} +# CHECK-NEXT: 2 12 1.00 * vpdpwssd (%rax){1to16}, %zmm1, %zmm2 {%k1} {z} +# CHECK-NEXT: 1 5 1.00 vpdpwssds %zmm0, %zmm1, %zmm2 +# CHECK-NEXT: 2 12 1.00 * vpdpwssds (%rax), %zmm1, %zmm2 +# CHECK-NEXT: 2 12 1.00 * vpdpwssds (%rax){1to16}, %zmm1, %zmm2 +# CHECK-NEXT: 1 5 1.00 vpdpwssds %zmm0, %zmm1, %zmm2 {%k1} +# CHECK-NEXT: 2 12 1.00 * vpdpwssds (%rax), %zmm1, %zmm2 {%k1} +# CHECK-NEXT: 2 12 1.00 * vpdpwssds (%rax){1to16}, %zmm1, %zmm2 {%k1} +# CHECK-NEXT: 1 5 1.00 vpdpwssds %zmm0, %zmm1, %zmm2 {%k1} {z} +# CHECK-NEXT: 2 12 1.00 * vpdpwssds (%rax), %zmm1, %zmm2 {%k1} {z} +# CHECK-NEXT: 2 12 1.00 * vpdpwssds (%rax){1to16}, %zmm1, %zmm2 {%k1} {z} + +# CHECK: Resources: +# CHECK-NEXT: [0] - SBDivider +# CHECK-NEXT: [1] - SBFPDivider +# CHECK-NEXT: [2] - SBPort0 +# CHECK-NEXT: [3] - SBPort1 +# CHECK-NEXT: [4] - SBPort4 +# CHECK-NEXT: [5] - SBPort5 +# CHECK-NEXT: [6.0] - SBPort23 +# CHECK-NEXT: [6.1] - SBPort23 + +# CHECK: Resource pressure per iteration: +# CHECK-NEXT: [0] [1] [2] [3] [4] [5] [6.0] [6.1] +# CHECK-NEXT: - - 36.00 - - - 12.00 12.00 + +# CHECK: Resource pressure by instruction: +# CHECK-NEXT: [0] [1] [2] [3] [4] [5] [6.0] [6.1] Instructions: +# CHECK-NEXT: - - 1.00 - - - - - vpdpbusd %zmm0, %zmm1, %zmm2 +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpbusd (%rax), %zmm1, %zmm2 +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpbusd (%rax){1to16}, %zmm1, %zmm2 +# CHECK-NEXT: - - 1.00 - - - - - vpdpbusd %zmm0, %zmm1, %zmm2 {%k1} +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpbusd (%rax), %zmm1, %zmm2 {%k1} +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpbusd (%rax){1to16}, %zmm1, %zmm2 {%k1} +# CHECK-NEXT: - - 1.00 - - - - - vpdpbusd %zmm0, %zmm1, %zmm2 {%k1} {z} +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpbusd (%rax), %zmm1, %zmm2 {%k1} {z} +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpbusd (%rax){1to16}, %zmm1, %zmm2 {%k1} {z} +# CHECK-NEXT: - - 1.00 - - - - - vpdpbusds %zmm0, %zmm1, %zmm2 +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpbusds (%rax), %zmm1, %zmm2 +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpbusds (%rax){1to16}, %zmm1, %zmm2 +# CHECK-NEXT: - - 1.00 - - - - - vpdpbusds %zmm0, %zmm1, %zmm2 {%k1} +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpbusds (%rax), %zmm1, %zmm2 {%k1} +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpbusds (%rax){1to16}, %zmm1, %zmm2 {%k1} +# CHECK-NEXT: - - 1.00 - - - - - vpdpbusds %zmm0, %zmm1, %zmm2 {%k1} {z} +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpbusds (%rax), %zmm1, %zmm2 {%k1} {z} +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpbusds (%rax){1to16}, %zmm1, %zmm2 {%k1} {z} +# CHECK-NEXT: - - 1.00 - - - - - vpdpwssd %zmm0, %zmm1, %zmm2 +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpwssd (%rax), %zmm1, %zmm2 +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpwssd (%rax){1to16}, %zmm1, %zmm2 +# CHECK-NEXT: - - 1.00 - - - - - vpdpwssd %zmm0, %zmm1, %zmm2 {%k1} +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpwssd (%rax), %zmm1, %zmm2 {%k1} +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpwssd (%rax){1to16}, %zmm1, %zmm2 {%k1} +# CHECK-NEXT: - - 1.00 - - - - - vpdpwssd %zmm0, %zmm1, %zmm2 {%k1} {z} +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpwssd (%rax), %zmm1, %zmm2 {%k1} {z} +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpwssd (%rax){1to16}, %zmm1, %zmm2 {%k1} {z} +# CHECK-NEXT: - - 1.00 - - - - - vpdpwssds %zmm0, %zmm1, %zmm2 +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpwssds (%rax), %zmm1, %zmm2 +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpwssds (%rax){1to16}, %zmm1, %zmm2 +# CHECK-NEXT: - - 1.00 - - - - - vpdpwssds %zmm0, %zmm1, %zmm2 {%k1} +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpwssds (%rax), %zmm1, %zmm2 {%k1} +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpwssds (%rax){1to16}, %zmm1, %zmm2 {%k1} +# CHECK-NEXT: - - 1.00 - - - - - vpdpwssds %zmm0, %zmm1, %zmm2 {%k1} {z} +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpwssds (%rax), %zmm1, %zmm2 {%k1} {z} +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpwssds (%rax){1to16}, %zmm1, %zmm2 {%k1} {z} diff --git a/llvm/test/tools/llvm-mca/X86/Generic/resources-avx512vnnivl.s b/llvm/test/tools/llvm-mca/X86/Generic/resources-avx512vnnivl.s new file mode 100644 index 000000000000..df2f4b81bede --- /dev/null +++ b/llvm/test/tools/llvm-mca/X86/Generic/resources-avx512vnnivl.s @@ -0,0 +1,253 @@ +# NOTE: Assertions have been autogenerated by utils/update_mca_test_checks.py +# RUN: llvm-mca -mtriple=x86_64-unknown-unknown -mcpu=x86-64 -instruction-tables < %s | FileCheck %s + +vpdpbusd %xmm0, %xmm1, %xmm2 +vpdpbusd (%rax), %xmm1, %xmm2 +vpdpbusd (%rax){1to4}, %xmm1, %xmm2 +vpdpbusd %xmm0, %xmm1, %xmm2 {k1} +vpdpbusd (%rax), %xmm1, %xmm2 {k1} +vpdpbusd (%rax){1to4}, %xmm1, %xmm2 {k1} +vpdpbusd %xmm0, %xmm1, %xmm2 {z}{k1} +vpdpbusd (%rax), %xmm1, %xmm2 {z}{k1} +vpdpbusd (%rax){1to4}, %xmm1, %xmm2 {z}{k1} + +vpdpbusd %ymm0, %ymm1, %ymm2 +vpdpbusd (%rax), %ymm1, %ymm2 +vpdpbusd (%rax){1to8}, %ymm1, %ymm2 +vpdpbusd %ymm0, %ymm1, %ymm2 {k1} +vpdpbusd (%rax), %ymm1, %ymm2 {k1} +vpdpbusd (%rax){1to8}, %ymm1, %ymm2 {k1} +vpdpbusd %ymm0, %ymm1, %ymm2 {z}{k1} +vpdpbusd (%rax), %ymm1, %ymm2 {z}{k1} +vpdpbusd (%rax){1to8}, %ymm1, %ymm2 {z}{k1} + +vpdpbusds %xmm0, %xmm1, %xmm2 +vpdpbusds (%rax), %xmm1, %xmm2 +vpdpbusds (%rax){1to4}, %xmm1, %xmm2 +vpdpbusds %xmm0, %xmm1, %xmm2 {k1} +vpdpbusds (%rax), %xmm1, %xmm2 {k1} +vpdpbusds (%rax){1to4}, %xmm1, %xmm2 {k1} +vpdpbusds %xmm0, %xmm1, %xmm2 {z}{k1} +vpdpbusds (%rax), %xmm1, %xmm2 {z}{k1} +vpdpbusds (%rax){1to4}, %xmm1, %xmm2 {z}{k1} + +vpdpbusds %ymm0, %ymm1, %ymm2 +vpdpbusds (%rax), %ymm1, %ymm2 +vpdpbusds (%rax){1to8}, %ymm1, %ymm2 +vpdpbusds %ymm0, %ymm1, %ymm2 {k1} +vpdpbusds (%rax), %ymm1, %ymm2 {k1} +vpdpbusds (%rax){1to8}, %ymm1, %ymm2 {k1} +vpdpbusds %ymm0, %ymm1, %ymm2 {z}{k1} +vpdpbusds (%rax), %ymm1, %ymm2 {z}{k1} +vpdpbusds (%rax){1to8}, %ymm1, %ymm2 {z}{k1} + +vpdpwssd %xmm0, %xmm1, %xmm2 +vpdpwssd (%rax), %xmm1, %xmm2 +vpdpwssd (%rax){1to4}, %xmm1, %xmm2 +vpdpwssd %xmm0, %xmm1, %xmm2 {k1} +vpdpwssd (%rax), %xmm1, %xmm2 {k1} +vpdpwssd (%rax){1to4}, %xmm1, %xmm2 {k1} +vpdpwssd %xmm0, %xmm1, %xmm2 {z}{k1} +vpdpwssd (%rax), %xmm1, %xmm2 {z}{k1} +vpdpwssd (%rax){1to4}, %xmm1, %xmm2 {z}{k1} + +vpdpwssd %ymm0, %ymm1, %ymm2 +vpdpwssd (%rax), %ymm1, %ymm2 +vpdpwssd (%rax){1to8}, %ymm1, %ymm2 +vpdpwssd %ymm0, %ymm1, %ymm2 {k1} +vpdpwssd (%rax), %ymm1, %ymm2 {k1} +vpdpwssd (%rax){1to8}, %ymm1, %ymm2 {k1} +vpdpwssd %ymm0, %ymm1, %ymm2 {z}{k1} +vpdpwssd (%rax), %ymm1, %ymm2 {z}{k1} +vpdpwssd (%rax){1to8}, %ymm1, %ymm2 {z}{k1} + +vpdpwssds %xmm0, %xmm1, %xmm2 +vpdpwssds (%rax), %xmm1, %xmm2 +vpdpwssds (%rax){1to4}, %xmm1, %xmm2 +vpdpwssds %xmm0, %xmm1, %xmm2 {k1} +vpdpwssds (%rax), %xmm1, %xmm2 {k1} +vpdpwssds (%rax){1to4}, %xmm1, %xmm2 {k1} +vpdpwssds %xmm0, %xmm1, %xmm2 {z}{k1} +vpdpwssds (%rax), %xmm1, %xmm2 {z}{k1} +vpdpwssds (%rax){1to4}, %xmm1, %xmm2 {z}{k1} + +vpdpwssds %ymm0, %ymm1, %ymm2 +vpdpwssds (%rax), %ymm1, %ymm2 +vpdpwssds (%rax){1to8}, %ymm1, %ymm2 +vpdpwssds %ymm0, %ymm1, %ymm2 {k1} +vpdpwssds (%rax), %ymm1, %ymm2 {k1} +vpdpwssds (%rax){1to8}, %ymm1, %ymm2 {k1} +vpdpwssds %ymm0, %ymm1, %ymm2 {z}{k1} +vpdpwssds (%rax), %ymm1, %ymm2 {z}{k1} +vpdpwssds (%rax){1to8}, %ymm1, %ymm2 {z}{k1} + +# CHECK: Instruction Info: +# CHECK-NEXT: [1]: #uOps +# CHECK-NEXT: [2]: Latency +# CHECK-NEXT: [3]: RThroughput +# CHECK-NEXT: [4]: MayLoad +# CHECK-NEXT: [5]: MayStore +# CHECK-NEXT: [6]: HasSideEffects (U) + +# CHECK: [1] [2] [3] [4] [5] [6] Instructions: +# CHECK-NEXT: 1 5 1.00 vpdpbusd %xmm0, %xmm1, %xmm2 +# CHECK-NEXT: 2 11 1.00 * vpdpbusd (%rax), %xmm1, %xmm2 +# CHECK-NEXT: 2 11 1.00 * vpdpbusd (%rax){1to4}, %xmm1, %xmm2 +# CHECK-NEXT: 1 5 1.00 vpdpbusd %xmm0, %xmm1, %xmm2 {%k1} +# CHECK-NEXT: 2 11 1.00 * vpdpbusd (%rax), %xmm1, %xmm2 {%k1} +# CHECK-NEXT: 2 11 1.00 * vpdpbusd (%rax){1to4}, %xmm1, %xmm2 {%k1} +# CHECK-NEXT: 1 5 1.00 vpdpbusd %xmm0, %xmm1, %xmm2 {%k1} {z} +# CHECK-NEXT: 2 11 1.00 * vpdpbusd (%rax), %xmm1, %xmm2 {%k1} {z} +# CHECK-NEXT: 2 11 1.00 * vpdpbusd (%rax){1to4}, %xmm1, %xmm2 {%k1} {z} +# CHECK-NEXT: 1 5 1.00 vpdpbusd %ymm0, %ymm1, %ymm2 +# CHECK-NEXT: 2 12 1.00 * vpdpbusd (%rax), %ymm1, %ymm2 +# CHECK-NEXT: 2 12 1.00 * vpdpbusd (%rax){1to8}, %ymm1, %ymm2 +# CHECK-NEXT: 1 5 1.00 vpdpbusd %ymm0, %ymm1, %ymm2 {%k1} +# CHECK-NEXT: 2 12 1.00 * vpdpbusd (%rax), %ymm1, %ymm2 {%k1} +# CHECK-NEXT: 2 12 1.00 * vpdpbusd (%rax){1to8}, %ymm1, %ymm2 {%k1} +# CHECK-NEXT: 1 5 1.00 vpdpbusd %ymm0, %ymm1, %ymm2 {%k1} {z} +# CHECK-NEXT: 2 12 1.00 * vpdpbusd (%rax), %ymm1, %ymm2 {%k1} {z} +# CHECK-NEXT: 2 12 1.00 * vpdpbusd (%rax){1to8}, %ymm1, %ymm2 {%k1} {z} +# CHECK-NEXT: 1 5 1.00 vpdpbusds %xmm0, %xmm1, %xmm2 +# CHECK-NEXT: 2 11 1.00 * vpdpbusds (%rax), %xmm1, %xmm2 +# CHECK-NEXT: 2 11 1.00 * vpdpbusds (%rax){1to4}, %xmm1, %xmm2 +# CHECK-NEXT: 1 5 1.00 vpdpbusds %xmm0, %xmm1, %xmm2 {%k1} +# CHECK-NEXT: 2 11 1.00 * vpdpbusds (%rax), %xmm1, %xmm2 {%k1} +# CHECK-NEXT: 2 11 1.00 * vpdpbusds (%rax){1to4}, %xmm1, %xmm2 {%k1} +# CHECK-NEXT: 1 5 1.00 vpdpbusds %xmm0, %xmm1, %xmm2 {%k1} {z} +# CHECK-NEXT: 2 11 1.00 * vpdpbusds (%rax), %xmm1, %xmm2 {%k1} {z} +# CHECK-NEXT: 2 11 1.00 * vpdpbusds (%rax){1to4}, %xmm1, %xmm2 {%k1} {z} +# CHECK-NEXT: 1 5 1.00 vpdpbusds %ymm0, %ymm1, %ymm2 +# CHECK-NEXT: 2 12 1.00 * vpdpbusds (%rax), %ymm1, %ymm2 +# CHECK-NEXT: 2 12 1.00 * vpdpbusds (%rax){1to8}, %ymm1, %ymm2 +# CHECK-NEXT: 1 5 1.00 vpdpbusds %ymm0, %ymm1, %ymm2 {%k1} +# CHECK-NEXT: 2 12 1.00 * vpdpbusds (%rax), %ymm1, %ymm2 {%k1} +# CHECK-NEXT: 2 12 1.00 * vpdpbusds (%rax){1to8}, %ymm1, %ymm2 {%k1} +# CHECK-NEXT: 1 5 1.00 vpdpbusds %ymm0, %ymm1, %ymm2 {%k1} {z} +# CHECK-NEXT: 2 12 1.00 * vpdpbusds (%rax), %ymm1, %ymm2 {%k1} {z} +# CHECK-NEXT: 2 12 1.00 * vpdpbusds (%rax){1to8}, %ymm1, %ymm2 {%k1} {z} +# CHECK-NEXT: 1 5 1.00 vpdpwssd %xmm0, %xmm1, %xmm2 +# CHECK-NEXT: 2 11 1.00 * vpdpwssd (%rax), %xmm1, %xmm2 +# CHECK-NEXT: 2 11 1.00 * vpdpwssd (%rax){1to4}, %xmm1, %xmm2 +# CHECK-NEXT: 1 5 1.00 vpdpwssd %xmm0, %xmm1, %xmm2 {%k1} +# CHECK-NEXT: 2 11 1.00 * vpdpwssd (%rax), %xmm1, %xmm2 {%k1} +# CHECK-NEXT: 2 11 1.00 * vpdpwssd (%rax){1to4}, %xmm1, %xmm2 {%k1} +# CHECK-NEXT: 1 5 1.00 vpdpwssd %xmm0, %xmm1, %xmm2 {%k1} {z} +# CHECK-NEXT: 2 11 1.00 * vpdpwssd (%rax), %xmm1, %xmm2 {%k1} {z} +# CHECK-NEXT: 2 11 1.00 * vpdpwssd (%rax){1to4}, %xmm1, %xmm2 {%k1} {z} +# CHECK-NEXT: 1 5 1.00 vpdpwssd %ymm0, %ymm1, %ymm2 +# CHECK-NEXT: 2 12 1.00 * vpdpwssd (%rax), %ymm1, %ymm2 +# CHECK-NEXT: 2 12 1.00 * vpdpwssd (%rax){1to8}, %ymm1, %ymm2 +# CHECK-NEXT: 1 5 1.00 vpdpwssd %ymm0, %ymm1, %ymm2 {%k1} +# CHECK-NEXT: 2 12 1.00 * vpdpwssd (%rax), %ymm1, %ymm2 {%k1} +# CHECK-NEXT: 2 12 1.00 * vpdpwssd (%rax){1to8}, %ymm1, %ymm2 {%k1} +# CHECK-NEXT: 1 5 1.00 vpdpwssd %ymm0, %ymm1, %ymm2 {%k1} {z} +# CHECK-NEXT: 2 12 1.00 * vpdpwssd (%rax), %ymm1, %ymm2 {%k1} {z} +# CHECK-NEXT: 2 12 1.00 * vpdpwssd (%rax){1to8}, %ymm1, %ymm2 {%k1} {z} +# CHECK-NEXT: 1 5 1.00 vpdpwssds %xmm0, %xmm1, %xmm2 +# CHECK-NEXT: 2 11 1.00 * vpdpwssds (%rax), %xmm1, %xmm2 +# CHECK-NEXT: 2 11 1.00 * vpdpwssds (%rax){1to4}, %xmm1, %xmm2 +# CHECK-NEXT: 1 5 1.00 vpdpwssds %xmm0, %xmm1, %xmm2 {%k1} +# CHECK-NEXT: 2 11 1.00 * vpdpwssds (%rax), %xmm1, %xmm2 {%k1} +# CHECK-NEXT: 2 11 1.00 * vpdpwssds (%rax){1to4}, %xmm1, %xmm2 {%k1} +# CHECK-NEXT: 1 5 1.00 vpdpwssds %xmm0, %xmm1, %xmm2 {%k1} {z} +# CHECK-NEXT: 2 11 1.00 * vpdpwssds (%rax), %xmm1, %xmm2 {%k1} {z} +# CHECK-NEXT: 2 11 1.00 * vpdpwssds (%rax){1to4}, %xmm1, %xmm2 {%k1} {z} +# CHECK-NEXT: 1 5 1.00 vpdpwssds %ymm0, %ymm1, %ymm2 +# CHECK-NEXT: 2 12 1.00 * vpdpwssds (%rax), %ymm1, %ymm2 +# CHECK-NEXT: 2 12 1.00 * vpdpwssds (%rax){1to8}, %ymm1, %ymm2 +# CHECK-NEXT: 1 5 1.00 vpdpwssds %ymm0, %ymm1, %ymm2 {%k1} +# CHECK-NEXT: 2 12 1.00 * vpdpwssds (%rax), %ymm1, %ymm2 {%k1} +# CHECK-NEXT: 2 12 1.00 * vpdpwssds (%rax){1to8}, %ymm1, %ymm2 {%k1} +# CHECK-NEXT: 1 5 1.00 vpdpwssds %ymm0, %ymm1, %ymm2 {%k1} {z} +# CHECK-NEXT: 2 12 1.00 * vpdpwssds (%rax), %ymm1, %ymm2 {%k1} {z} +# CHECK-NEXT: 2 12 1.00 * vpdpwssds (%rax){1to8}, %ymm1, %ymm2 {%k1} {z} + +# CHECK: Resources: +# CHECK-NEXT: [0] - SBDivider +# CHECK-NEXT: [1] - SBFPDivider +# CHECK-NEXT: [2] - SBPort0 +# CHECK-NEXT: [3] - SBPort1 +# CHECK-NEXT: [4] - SBPort4 +# CHECK-NEXT: [5] - SBPort5 +# CHECK-NEXT: [6.0] - SBPort23 +# CHECK-NEXT: [6.1] - SBPort23 + +# CHECK: Resource pressure per iteration: +# CHECK-NEXT: [0] [1] [2] [3] [4] [5] [6.0] [6.1] +# CHECK-NEXT: - - 72.00 - - - 24.00 24.00 + +# CHECK: Resource pressure by instruction: +# CHECK-NEXT: [0] [1] [2] [3] [4] [5] [6.0] [6.1] Instructions: +# CHECK-NEXT: - - 1.00 - - - - - vpdpbusd %xmm0, %xmm1, %xmm2 +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpbusd (%rax), %xmm1, %xmm2 +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpbusd (%rax){1to4}, %xmm1, %xmm2 +# CHECK-NEXT: - - 1.00 - - - - - vpdpbusd %xmm0, %xmm1, %xmm2 {%k1} +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpbusd (%rax), %xmm1, %xmm2 {%k1} +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpbusd (%rax){1to4}, %xmm1, %xmm2 {%k1} +# CHECK-NEXT: - - 1.00 - - - - - vpdpbusd %xmm0, %xmm1, %xmm2 {%k1} {z} +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpbusd (%rax), %xmm1, %xmm2 {%k1} {z} +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpbusd (%rax){1to4}, %xmm1, %xmm2 {%k1} {z} +# CHECK-NEXT: - - 1.00 - - - - - vpdpbusd %ymm0, %ymm1, %ymm2 +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpbusd (%rax), %ymm1, %ymm2 +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpbusd (%rax){1to8}, %ymm1, %ymm2 +# CHECK-NEXT: - - 1.00 - - - - - vpdpbusd %ymm0, %ymm1, %ymm2 {%k1} +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpbusd (%rax), %ymm1, %ymm2 {%k1} +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpbusd (%rax){1to8}, %ymm1, %ymm2 {%k1} +# CHECK-NEXT: - - 1.00 - - - - - vpdpbusd %ymm0, %ymm1, %ymm2 {%k1} {z} +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpbusd (%rax), %ymm1, %ymm2 {%k1} {z} +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpbusd (%rax){1to8}, %ymm1, %ymm2 {%k1} {z} +# CHECK-NEXT: - - 1.00 - - - - - vpdpbusds %xmm0, %xmm1, %xmm2 +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpbusds (%rax), %xmm1, %xmm2 +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpbusds (%rax){1to4}, %xmm1, %xmm2 +# CHECK-NEXT: - - 1.00 - - - - - vpdpbusds %xmm0, %xmm1, %xmm2 {%k1} +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpbusds (%rax), %xmm1, %xmm2 {%k1} +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpbusds (%rax){1to4}, %xmm1, %xmm2 {%k1} +# CHECK-NEXT: - - 1.00 - - - - - vpdpbusds %xmm0, %xmm1, %xmm2 {%k1} {z} +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpbusds (%rax), %xmm1, %xmm2 {%k1} {z} +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpbusds (%rax){1to4}, %xmm1, %xmm2 {%k1} {z} +# CHECK-NEXT: - - 1.00 - - - - - vpdpbusds %ymm0, %ymm1, %ymm2 +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpbusds (%rax), %ymm1, %ymm2 +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpbusds (%rax){1to8}, %ymm1, %ymm2 +# CHECK-NEXT: - - 1.00 - - - - - vpdpbusds %ymm0, %ymm1, %ymm2 {%k1} +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpbusds (%rax), %ymm1, %ymm2 {%k1} +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpbusds (%rax){1to8}, %ymm1, %ymm2 {%k1} +# CHECK-NEXT: - - 1.00 - - - - - vpdpbusds %ymm0, %ymm1, %ymm2 {%k1} {z} +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpbusds (%rax), %ymm1, %ymm2 {%k1} {z} +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpbusds (%rax){1to8}, %ymm1, %ymm2 {%k1} {z} +# CHECK-NEXT: - - 1.00 - - - - - vpdpwssd %xmm0, %xmm1, %xmm2 +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpwssd (%rax), %xmm1, %xmm2 +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpwssd (%rax){1to4}, %xmm1, %xmm2 +# CHECK-NEXT: - - 1.00 - - - - - vpdpwssd %xmm0, %xmm1, %xmm2 {%k1} +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpwssd (%rax), %xmm1, %xmm2 {%k1} +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpwssd (%rax){1to4}, %xmm1, %xmm2 {%k1} +# CHECK-NEXT: - - 1.00 - - - - - vpdpwssd %xmm0, %xmm1, %xmm2 {%k1} {z} +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpwssd (%rax), %xmm1, %xmm2 {%k1} {z} +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpwssd (%rax){1to4}, %xmm1, %xmm2 {%k1} {z} +# CHECK-NEXT: - - 1.00 - - - - - vpdpwssd %ymm0, %ymm1, %ymm2 +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpwssd (%rax), %ymm1, %ymm2 +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpwssd (%rax){1to8}, %ymm1, %ymm2 +# CHECK-NEXT: - - 1.00 - - - - - vpdpwssd %ymm0, %ymm1, %ymm2 {%k1} +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpwssd (%rax), %ymm1, %ymm2 {%k1} +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpwssd (%rax){1to8}, %ymm1, %ymm2 {%k1} +# CHECK-NEXT: - - 1.00 - - - - - vpdpwssd %ymm0, %ymm1, %ymm2 {%k1} {z} +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpwssd (%rax), %ymm1, %ymm2 {%k1} {z} +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpwssd (%rax){1to8}, %ymm1, %ymm2 {%k1} {z} +# CHECK-NEXT: - - 1.00 - - - - - vpdpwssds %xmm0, %xmm1, %xmm2 +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpwssds (%rax), %xmm1, %xmm2 +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpwssds (%rax){1to4}, %xmm1, %xmm2 +# CHECK-NEXT: - - 1.00 - - - - - vpdpwssds %xmm0, %xmm1, %xmm2 {%k1} +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpwssds (%rax), %xmm1, %xmm2 {%k1} +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpwssds (%rax){1to4}, %xmm1, %xmm2 {%k1} +# CHECK-NEXT: - - 1.00 - - - - - vpdpwssds %xmm0, %xmm1, %xmm2 {%k1} {z} +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpwssds (%rax), %xmm1, %xmm2 {%k1} {z} +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpwssds (%rax){1to4}, %xmm1, %xmm2 {%k1} {z} +# CHECK-NEXT: - - 1.00 - - - - - vpdpwssds %ymm0, %ymm1, %ymm2 +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpwssds (%rax), %ymm1, %ymm2 +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpwssds (%rax){1to8}, %ymm1, %ymm2 +# CHECK-NEXT: - - 1.00 - - - - - vpdpwssds %ymm0, %ymm1, %ymm2 {%k1} +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpwssds (%rax), %ymm1, %ymm2 {%k1} +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpwssds (%rax){1to8}, %ymm1, %ymm2 {%k1} +# CHECK-NEXT: - - 1.00 - - - - - vpdpwssds %ymm0, %ymm1, %ymm2 {%k1} {z} +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpwssds (%rax), %ymm1, %ymm2 {%k1} {z} +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpwssds (%rax){1to8}, %ymm1, %ymm2 {%k1} {z} diff --git a/llvm/test/tools/llvm-mca/X86/Generic/resources-avxvnni.s b/llvm/test/tools/llvm-mca/X86/Generic/resources-avxvnni.s new file mode 100644 index 000000000000..cb1f166e41d7 --- /dev/null +++ b/llvm/test/tools/llvm-mca/X86/Generic/resources-avxvnni.s @@ -0,0 +1,85 @@ +# NOTE: Assertions have been autogenerated by utils/update_mca_test_checks.py +# RUN: llvm-mca -mtriple=x86_64-unknown-unknown -mcpu=x86-64 -instruction-tables < %s | FileCheck %s + +vpdpbusd %xmm0, %xmm1, %xmm2 +vpdpbusd (%rax), %xmm1, %xmm2 + +vpdpbusd %ymm0, %ymm1, %ymm2 +vpdpbusd (%rax), %ymm1, %ymm2 + +vpdpbusds %xmm0, %xmm1, %xmm2 +vpdpbusds (%rax), %xmm1, %xmm2 + +vpdpbusds %ymm0, %ymm1, %ymm2 +vpdpbusds (%rax), %ymm1, %ymm2 + +vpdpwssd %xmm0, %xmm1, %xmm2 +vpdpwssd (%rax), %xmm1, %xmm2 + +vpdpwssd %ymm0, %ymm1, %ymm2 +vpdpwssd (%rax), %ymm1, %ymm2 + +vpdpwssds %xmm0, %xmm1, %xmm2 +vpdpwssds (%rax), %xmm1, %xmm2 + +vpdpwssds %ymm0, %ymm1, %ymm2 +vpdpwssds (%rax), %ymm1, %ymm2 + +# CHECK: Instruction Info: +# CHECK-NEXT: [1]: #uOps +# CHECK-NEXT: [2]: Latency +# CHECK-NEXT: [3]: RThroughput +# CHECK-NEXT: [4]: MayLoad +# CHECK-NEXT: [5]: MayStore +# CHECK-NEXT: [6]: HasSideEffects (U) + +# CHECK: [1] [2] [3] [4] [5] [6] Instructions: +# CHECK-NEXT: 1 5 1.00 vpdpbusd %xmm0, %xmm1, %xmm2 +# CHECK-NEXT: 2 11 1.00 * vpdpbusd (%rax), %xmm1, %xmm2 +# CHECK-NEXT: 1 5 1.00 vpdpbusd %ymm0, %ymm1, %ymm2 +# CHECK-NEXT: 2 12 1.00 * vpdpbusd (%rax), %ymm1, %ymm2 +# CHECK-NEXT: 1 5 1.00 vpdpbusds %xmm0, %xmm1, %xmm2 +# CHECK-NEXT: 2 11 1.00 * vpdpbusds (%rax), %xmm1, %xmm2 +# CHECK-NEXT: 1 5 1.00 vpdpbusds %ymm0, %ymm1, %ymm2 +# CHECK-NEXT: 2 12 1.00 * vpdpbusds (%rax), %ymm1, %ymm2 +# CHECK-NEXT: 1 5 1.00 vpdpwssd %xmm0, %xmm1, %xmm2 +# CHECK-NEXT: 2 11 1.00 * vpdpwssd (%rax), %xmm1, %xmm2 +# CHECK-NEXT: 1 5 1.00 vpdpwssd %ymm0, %ymm1, %ymm2 +# CHECK-NEXT: 2 12 1.00 * vpdpwssd (%rax), %ymm1, %ymm2 +# CHECK-NEXT: 1 5 1.00 vpdpwssds %xmm0, %xmm1, %xmm2 +# CHECK-NEXT: 2 11 1.00 * vpdpwssds (%rax), %xmm1, %xmm2 +# CHECK-NEXT: 1 5 1.00 vpdpwssds %ymm0, %ymm1, %ymm2 +# CHECK-NEXT: 2 12 1.00 * vpdpwssds (%rax), %ymm1, %ymm2 + +# CHECK: Resources: +# CHECK-NEXT: [0] - SBDivider +# CHECK-NEXT: [1] - SBFPDivider +# CHECK-NEXT: [2] - SBPort0 +# CHECK-NEXT: [3] - SBPort1 +# CHECK-NEXT: [4] - SBPort4 +# CHECK-NEXT: [5] - SBPort5 +# CHECK-NEXT: [6.0] - SBPort23 +# CHECK-NEXT: [6.1] - SBPort23 + +# CHECK: Resource pressure per iteration: +# CHECK-NEXT: [0] [1] [2] [3] [4] [5] [6.0] [6.1] +# CHECK-NEXT: - - 16.00 - - - 4.00 4.00 + +# CHECK: Resource pressure by instruction: +# CHECK-NEXT: [0] [1] [2] [3] [4] [5] [6.0] [6.1] Instructions: +# CHECK-NEXT: - - 1.00 - - - - - vpdpbusd %xmm0, %xmm1, %xmm2 +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpbusd (%rax), %xmm1, %xmm2 +# CHECK-NEXT: - - 1.00 - - - - - vpdpbusd %ymm0, %ymm1, %ymm2 +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpbusd (%rax), %ymm1, %ymm2 +# CHECK-NEXT: - - 1.00 - - - - - vpdpbusds %xmm0, %xmm1, %xmm2 +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpbusds (%rax), %xmm1, %xmm2 +# CHECK-NEXT: - - 1.00 - - - - - vpdpbusds %ymm0, %ymm1, %ymm2 +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpbusds (%rax), %ymm1, %ymm2 +# CHECK-NEXT: - - 1.00 - - - - - vpdpwssd %xmm0, %xmm1, %xmm2 +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpwssd (%rax), %xmm1, %xmm2 +# CHECK-NEXT: - - 1.00 - - - - - vpdpwssd %ymm0, %ymm1, %ymm2 +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpwssd (%rax), %ymm1, %ymm2 +# CHECK-NEXT: - - 1.00 - - - - - vpdpwssds %xmm0, %xmm1, %xmm2 +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpwssds (%rax), %xmm1, %xmm2 +# CHECK-NEXT: - - 1.00 - - - - - vpdpwssds %ymm0, %ymm1, %ymm2 +# CHECK-NEXT: - - 1.00 - - - 0.50 0.50 vpdpwssds (%rax), %ymm1, %ymm2 diff --git a/llvm/test/tools/llvm-mca/X86/IceLakeServer/resources-avx512vnni.s b/llvm/test/tools/llvm-mca/X86/IceLakeServer/resources-avx512vnni.s new file mode 100644 index 000000000000..1c24f350f877 --- /dev/null +++ b/llvm/test/tools/llvm-mca/X86/IceLakeServer/resources-avx512vnni.s @@ -0,0 +1,145 @@ +# NOTE: Assertions have been autogenerated by utils/update_mca_test_checks.py +# RUN: llvm-mca -mtriple=x86_64-unknown-unknown -mcpu=icelake-server -instruction-tables < %s | FileCheck %s + +vpdpbusd %zmm0, %zmm1, %zmm2 +vpdpbusd (%rax), %zmm1, %zmm2 +vpdpbusd (%rax){1to16}, %zmm1, %zmm2 +vpdpbusd %zmm0, %zmm1, %zmm2 {k1} +vpdpbusd (%rax), %zmm1, %zmm2 {k1} +vpdpbusd (%rax){1to16}, %zmm1, %zmm2 {k1} +vpdpbusd %zmm0, %zmm1, %zmm2 {z}{k1} +vpdpbusd (%rax), %zmm1, %zmm2 {z}{k1} +vpdpbusd (%rax){1to16}, %zmm1, %zmm2 {z}{k1} + +vpdpbusds %zmm0, %zmm1, %zmm2 +vpdpbusds (%rax), %zmm1, %zmm2 +vpdpbusds (%rax){1to16}, %zmm1, %zmm2 +vpdpbusds %zmm0, %zmm1, %zmm2 {k1} +vpdpbusds (%rax), %zmm1, %zmm2 {k1} +vpdpbusds (%rax){1to16}, %zmm1, %zmm2 {k1} +vpdpbusds %zmm0, %zmm1, %zmm2 {z}{k1} +vpdpbusds (%rax), %zmm1, %zmm2 {z}{k1} +vpdpbusds (%rax){1to16}, %zmm1, %zmm2 {z}{k1} + +vpdpwssd %zmm0, %zmm1, %zmm2 +vpdpwssd (%rax), %zmm1, %zmm2 +vpdpwssd (%rax){1to16}, %zmm1, %zmm2 +vpdpwssd %zmm0, %zmm1, %zmm2 {k1} +vpdpwssd (%rax), %zmm1, %zmm2 {k1} +vpdpwssd (%rax){1to16}, %zmm1, %zmm2 {k1} +vpdpwssd %zmm0, %zmm1, %zmm2 {z}{k1} +vpdpwssd (%rax), %zmm1, %zmm2 {z}{k1} +vpdpwssd (%rax){1to16}, %zmm1, %zmm2 {z}{k1} + +vpdpwssds %zmm0, %zmm1, %zmm2 +vpdpwssds (%rax), %zmm1, %zmm2 +vpdpwssds (%rax){1to16}, %zmm1, %zmm2 +vpdpwssds %zmm0, %zmm1, %zmm2 {k1} +vpdpwssds (%rax), %zmm1, %zmm2 {k1} +vpdpwssds (%rax){1to16}, %zmm1, %zmm2 {k1} +vpdpwssds %zmm0, %zmm1, %zmm2 {z}{k1} +vpdpwssds (%rax), %zmm1, %zmm2 {z}{k1} +vpdpwssds (%rax){1to16}, %zmm1, %zmm2 {z}{k1} + +# CHECK: Instruction Info: +# CHECK-NEXT: [1]: #uOps +# CHECK-NEXT: [2]: Latency +# CHECK-NEXT: [3]: RThroughput +# CHECK-NEXT: [4]: MayLoad +# CHECK-NEXT: [5]: MayStore +# CHECK-NEXT: [6]: HasSideEffects (U) + +# CHECK: [1] [2] [3] [4] [5] [6] Instructions: +# CHECK-NEXT: 1 5 0.50 vpdpbusd %zmm0, %zmm1, %zmm2 +# CHECK-NEXT: 2 12 0.50 * vpdpbusd (%rax), %zmm1, %zmm2 +# CHECK-NEXT: 2 12 0.50 * vpdpbusd (%rax){1to16}, %zmm1, %zmm2 +# CHECK-NEXT: 1 5 0.50 vpdpbusd %zmm0, %zmm1, %zmm2 {%k1} +# CHECK-NEXT: 2 12 0.50 * vpdpbusd (%rax), %zmm1, %zmm2 {%k1} +# CHECK-NEXT: 2 12 0.50 * vpdpbusd (%rax){1to16}, %zmm1, %zmm2 {%k1} +# CHECK-NEXT: 1 5 0.50 vpdpbusd %zmm0, %zmm1, %zmm2 {%k1} {z} +# CHECK-NEXT: 2 12 0.50 * vpdpbusd (%rax), %zmm1, %zmm2 {%k1} {z} +# CHECK-NEXT: 2 12 0.50 * vpdpbusd (%rax){1to16}, %zmm1, %zmm2 {%k1} {z} +# CHECK-NEXT: 1 5 0.50 vpdpbusds %zmm0, %zmm1, %zmm2 +# CHECK-NEXT: 2 12 0.50 * vpdpbusds (%rax), %zmm1, %zmm2 +# CHECK-NEXT: 2 12 0.50 * vpdpbusds (%rax){1to16}, %zmm1, %zmm2 +# CHECK-NEXT: 1 5 0.50 vpdpbusds %zmm0, %zmm1, %zmm2 {%k1} +# CHECK-NEXT: 2 12 0.50 * vpdpbusds (%rax), %zmm1, %zmm2 {%k1} +# CHECK-NEXT: 2 12 0.50 * vpdpbusds (%rax){1to16}, %zmm1, %zmm2 {%k1} +# CHECK-NEXT: 1 5 0.50 vpdpbusds %zmm0, %zmm1, %zmm2 {%k1} {z} +# CHECK-NEXT: 2 12 0.50 * vpdpbusds (%rax), %zmm1, %zmm2 {%k1} {z} +# CHECK-NEXT: 2 12 0.50 * vpdpbusds (%rax){1to16}, %zmm1, %zmm2 {%k1} {z} +# CHECK-NEXT: 1 5 0.50 vpdpwssd %zmm0, %zmm1, %zmm2 +# CHECK-NEXT: 2 12 0.50 * vpdpwssd (%rax), %zmm1, %zmm2 +# CHECK-NEXT: 2 12 0.50 * vpdpwssd (%rax){1to16}, %zmm1, %zmm2 +# CHECK-NEXT: 1 5 0.50 vpdpwssd %zmm0, %zmm1, %zmm2 {%k1} +# CHECK-NEXT: 2 12 0.50 * vpdpwssd (%rax), %zmm1, %zmm2 {%k1} +# CHECK-NEXT: 2 12 0.50 * vpdpwssd (%rax){1to16}, %zmm1, %zmm2 {%k1} +# CHECK-NEXT: 1 5 0.50 vpdpwssd %zmm0, %zmm1, %zmm2 {%k1} {z} +# CHECK-NEXT: 2 12 0.50 * vpdpwssd (%rax), %zmm1, %zmm2 {%k1} {z} +# CHECK-NEXT: 2 12 0.50 * vpdpwssd (%rax){1to16}, %zmm1, %zmm2 {%k1} {z} +# CHECK-NEXT: 1 5 0.50 vpdpwssds %zmm0, %zmm1, %zmm2 +# CHECK-NEXT: 2 12 0.50 * vpdpwssds (%rax), %zmm1, %zmm2 +# CHECK-NEXT: 2 12 0.50 * vpdpwssds (%rax){1to16}, %zmm1, %zmm2 +# CHECK-NEXT: 1 5 0.50 vpdpwssds %zmm0, %zmm1, %zmm2 {%k1} +# CHECK-NEXT: 2 12 0.50 * vpdpwssds (%rax), %zmm1, %zmm2 {%k1} +# CHECK-NEXT: 2 12 0.50 * vpdpwssds (%rax){1to16}, %zmm1, %zmm2 {%k1} +# CHECK-NEXT: 1 5 0.50 vpdpwssds %zmm0, %zmm1, %zmm2 {%k1} {z} +# CHECK-NEXT: 2 12 0.50 * vpdpwssds (%rax), %zmm1, %zmm2 {%k1} {z} +# CHECK-NEXT: 2 12 0.50 * vpdpwssds (%rax){1to16}, %zmm1, %zmm2 {%k1} {z} + +# CHECK: Resources: +# CHECK-NEXT: [0] - ICXDivider +# CHECK-NEXT: [1] - ICXFPDivider +# CHECK-NEXT: [2] - ICXPort0 +# CHECK-NEXT: [3] - ICXPort1 +# CHECK-NEXT: [4] - ICXPort2 +# CHECK-NEXT: [5] - ICXPort3 +# CHECK-NEXT: [6] - ICXPort4 +# CHECK-NEXT: [7] - ICXPort5 +# CHECK-NEXT: [8] - ICXPort6 +# CHECK-NEXT: [9] - ICXPort7 +# CHECK-NEXT: [10] - ICXPort8 +# CHECK-NEXT: [11] - ICXPort9 + +# CHECK: Resource pressure per iteration: +# CHECK-NEXT: [0] [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] +# CHECK-NEXT: - - 18.00 - 12.00 12.00 - 18.00 - - - - + +# CHECK: Resource pressure by instruction: +# CHECK-NEXT: [0] [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] Instructions: +# CHECK-NEXT: - - 0.50 - - - - 0.50 - - - - vpdpbusd %zmm0, %zmm1, %zmm2 +# CHECK-NEXT: - - 0.50 - 0.50 0.50 - 0.50 - - - - vpdpbusd (%rax), %zmm1, %zmm2 +# CHECK-NEXT: - - 0.50 - 0.50 0.50 - 0.50 - - - - vpdpbusd (%rax){1to16}, %zmm1, %zmm2 +# CHECK-NEXT: - - 0.50 - - - - 0.50 - - - - vpdpbusd %zmm0, %zmm1, %zmm2 {%k1} +# CHECK-NEXT: - - 0.50 - 0.50 0.50 - 0.50 - - - - vpdpbusd (%rax), %zmm1, %zmm2 {%k1} +# CHECK-NEXT: - - 0.50 - 0.50 0.50 - 0.50 - - - - vpdpbusd (%rax){1to16}, %zmm1, %zmm2 {%k1} +# CHECK-NEXT: - - 0.50 - - - - 0.50 - - - - vpdpbusd %zmm0, %zmm1, %zmm2 {%k1} {z} +# CHECK-NEXT: - - 0.50 - 0.50 0.50 - 0.50 - - - - vpdpbusd (%rax), %zmm1, %zmm2 {%k1} {z} +# CHECK-NEXT: - - 0.50 - 0.50 0.50 - 0.50 - - - - vpdpbusd (%rax){1to16}, %zmm1, %zmm2 {%k1} {z} +# CHECK-NEXT: - - 0.50 - - - - 0.50 - - - - vpdpbusds %zmm0, %zmm1, %zmm2 +# CHECK-NEXT: - - 0.50 - 0.50 0.50 - 0.50 - - - - vpdpbusds (%rax), %zmm1, %zmm2 +# CHECK-NEXT: - - 0.50 - 0.50 0.50 - 0.50 - - - - vpdpbusds (%rax){1to16}, %zmm1, %zmm2 +# CHECK-NEXT: - - 0.50 - - - - 0.50 - - - - vpdpbusds %zmm0, %zmm1, %zmm2 {%k1} +# CHECK-NEXT: - - 0.50 - 0.50 0.50 - 0.50 - - - - vpdpbusds (%rax), %zmm1, %zmm2 {%k1} +# CHECK-NEXT: - - 0.50 - 0.50 0.50 - 0.50 - - - - vpdpbusds (%rax){1to16}, %zmm1, %zmm2 {%k1} +# CHECK-NEXT: - - 0.50 - - - - 0.50 - - - - vpdpbusds %zmm0, %zmm1, %zmm2 {%k1} {z} +# CHECK-NEXT: - - 0.50 - 0.50 0.50 - 0.50 - - - - vpdpbusds (%rax), %zmm1, %zmm2 {%k1} {z} +# CHECK-NEXT: - - 0.50 - 0.50 0.50 - 0.50 - - - - vpdpbusds (%rax){1to16}, %zmm1, %zmm2 {%k1} {z} +# CHECK-NEXT: - - 0.50 - - - - 0.50 - - - - vpdpwssd %zmm0, %zmm1, %zmm2 +# CHECK-NEXT: - - 0.50 - 0.50 0.50 - 0.50 - - - - vpdpwssd (%rax), %zmm1, %zmm2 +# CHECK-NEXT: - - 0.50 - 0.50 0.50 - 0.50 - - - - vpdpwssd (%rax){1to16}, %zmm1, %zmm2 +# CHECK-NEXT: - - 0.50 - - - - 0.50 - - - - vpdpwssd %zmm0, %zmm1, %zmm2 {%k1} +# CHECK-NEXT: - - 0.50 - 0.50 0.50 - 0.50 - - - - vpdpwssd (%rax), %zmm1, %zmm2 {%k1} +# CHECK-NEXT: - - 0.50 - 0.50 0.50 - 0.50 - - - - vpdpwssd (%rax){1to16}, %zmm1, %zmm2 {%k1} +# CHECK-NEXT: - - 0.50 - - - - 0.50 - - - - vpdpwssd %zmm0, %zmm1, %zmm2 {%k1} {z} +# CHECK-NEXT: - - 0.50 - 0.50 0.50 - 0.50 - - - - vpdpwssd (%rax), %zmm1, %zmm2 {%k1} {z} +# CHECK-NEXT: - - 0.50 - 0.50 0.50 - 0.50 - - - - vpdpwssd (%rax){1to16}, %zmm1, %zmm2 {%k1} {z} +# CHECK-NEXT: - - 0.50 - - - - 0.50 - - - - vpdpwssds %zmm0, %zmm1, %zmm2 +# CHECK-NEXT: - - 0.50 - 0.50 0.50 - 0.50 - - - - vpdpwssds (%rax), %zmm1, %zmm2 +# CHECK-NEXT: - - 0.50 - 0.50 0.50 - 0.50 - - - - vpdpwssds (%rax){1to16}, %zmm1, %zmm2 +# CHECK-NEXT: - - 0.50 - - - - 0.50 - - - - vpdpwssds %zmm0, %zmm1, %zmm2 {%k1} +# CHECK-NEXT: - - 0.50 - 0.50 0.50 - 0.50 - - - - vpdpwssds (%rax), %zmm1, %zmm2 {%k1} +# CHECK-NEXT: - - 0.50 - 0.50 0.50 - 0.50 - - - - vpdpwssds (%rax){1to16}, %zmm1, %zmm2 {%k1} +# CHECK-NEXT: - - 0.50 - - - - 0.50 - - - - vpdpwssds %zmm0, %zmm1, %zmm2 {%k1} {z} +# CHECK-NEXT: - - 0.50 - 0.50 0.50 - 0.50 - - - - vpdpwssds (%rax), %zmm1, %zmm2 {%k1} {z} +# CHECK-NEXT: - - 0.50 - 0.50 0.50 - 0.50 - - - - vpdpwssds (%rax){1to16}, %zmm1, %zmm2 {%k1} {z} diff --git a/llvm/test/tools/llvm-mca/X86/IceLakeServer/resources-avx512vnnivl.s b/llvm/test/tools/llvm-mca/X86/IceLakeServer/resources-avx512vnnivl.s new file mode 100644 index 000000000000..910fac829ea8 --- /dev/null +++ b/llvm/test/tools/llvm-mca/X86/IceLakeServer/resources-avx512vnnivl.s @@ -0,0 +1,257 @@ +# NOTE: Assertions have been autogenerated by utils/update_mca_test_checks.py +# RUN: llvm-mca -mtriple=x86_64-unknown-unknown -mcpu=icelake-server -instruction-tables < %s | FileCheck %s + +vpdpbusd %xmm0, %xmm1, %xmm2 +vpdpbusd (%rax), %xmm1, %xmm2 +vpdpbusd (%rax){1to4}, %xmm1, %xmm2 +vpdpbusd %xmm0, %xmm1, %xmm2 {k1} +vpdpbusd (%rax), %xmm1, %xmm2 {k1} +vpdpbusd (%rax){1to4}, %xmm1, %xmm2 {k1} +vpdpbusd %xmm0, %xmm1, %xmm2 {z}{k1} +vpdpbusd (%rax), %xmm1, %xmm2 {z}{k1} +vpdpbusd (%rax){1to4}, %xmm1, %xmm2 {z}{k1} + +vpdpbusd %ymm0, %ymm1, %ymm2 +vpdpbusd (%rax), %ymm1, %ymm2 +vpdpbusd (%rax){1to8}, %ymm1, %ymm2 +vpdpbusd %ymm0, %ymm1, %ymm2 {k1} +vpdpbusd (%rax), %ymm1, %ymm2 {k1} +vpdpbusd (%rax){1to8}, %ymm1, %ymm2 {k1} +vpdpbusd %ymm0, %ymm1, %ymm2 {z}{k1} +vpdpbusd (%rax), %ymm1, %ymm2 {z}{k1} +vpdpbusd (%rax){1to8}, %ymm1, %ymm2 {z}{k1} + +vpdpbusds %xmm0, %xmm1, %xmm2 +vpdpbusds (%rax), %xmm1, %xmm2 +vpdpbusds (%rax){1to4}, %xmm1, %xmm2 +vpdpbusds %xmm0, %xmm1, %xmm2 {k1} +vpdpbusds (%rax), %xmm1, %xmm2 {k1} +vpdpbusds (%rax){1to4}, %xmm1, %xmm2 {k1} +vpdpbusds %xmm0, %xmm1, %xmm2 {z}{k1} +vpdpbusds (%rax), %xmm1, %xmm2 {z}{k1} +vpdpbusds (%rax){1to4}, %xmm1, %xmm2 {z}{k1} + +vpdpbusds %ymm0, %ymm1, %ymm2 +vpdpbusds (%rax), %ymm1, %ymm2 +vpdpbusds (%rax){1to8}, %ymm1, %ymm2 +vpdpbusds %ymm0, %ymm1, %ymm2 {k1} +vpdpbusds (%rax), %ymm1, %ymm2 {k1} +vpdpbusds (%rax){1to8}, %ymm1, %ymm2 {k1} +vpdpbusds %ymm0, %ymm1, %ymm2 {z}{k1} +vpdpbusds (%rax), %ymm1, %ymm2 {z}{k1} +vpdpbusds (%rax){1to8}, %ymm1, %ymm2 {z}{k1} + +vpdpwssd %xmm0, %xmm1, %xmm2 +vpdpwssd (%rax), %xmm1, %xmm2 +vpdpwssd (%rax){1to4}, %xmm1, %xmm2 +vpdpwssd %xmm0, %xmm1, %xmm2 {k1} +vpdpwssd (%rax), %xmm1, %xmm2 {k1} +vpdpwssd (%rax){1to4}, %xmm1, %xmm2 {k1} +vpdpwssd %xmm0, %xmm1, %xmm2 {z}{k1} +vpdpwssd (%rax), %xmm1, %xmm2 {z}{k1} +vpdpwssd (%rax){1to4}, %xmm1, %xmm2 {z}{k1} + +vpdpwssd %ymm0, %ymm1, %ymm2 +vpdpwssd (%rax), %ymm1, %ymm2 +vpdpwssd (%rax){1to8}, %ymm1, %ymm2 +vpdpwssd %ymm0, %ymm1, %ymm2 {k1} +vpdpwssd (%rax), %ymm1, %ymm2 {k1} +vpdpwssd (%rax){1to8}, %ymm1, %ymm2 {k1} +vpdpwssd %ymm0, %ymm1, %ymm2 {z}{k1} +vpdpwssd (%rax), %ymm1, %ymm2 {z}{k1} +vpdpwssd (%rax){1to8}, %ymm1, %ymm2 {z}{k1} + +vpdpwssds %xmm0, %xmm1, %xmm2 +vpdpwssds (%rax), %xmm1, %xmm2 +vpdpwssds (%rax){1to4}, %xmm1, %xmm2 +vpdpwssds %xmm0, %xmm1, %xmm2 {k1} +vpdpwssds (%rax), %xmm1, %xmm2 {k1} +vpdpwssds (%rax){1to4}, %xmm1, %xmm2 {k1} +vpdpwssds %xmm0, %xmm1, %xmm2 {z}{k1} +vpdpwssds (%rax), %xmm1, %xmm2 {z}{k1} +vpdpwssds (%rax){1to4}, %xmm1, %xmm2 {z}{k1} + +vpdpwssds %ymm0, %ymm1, %ymm2 +vpdpwssds (%rax), %ymm1, %ymm2 +vpdpwssds (%rax){1to8}, %ymm1, %ymm2 +vpdpwssds %ymm0, %ymm1, %ymm2 {k1} +vpdpwssds (%rax), %ymm1, %ymm2 {k1} +vpdpwssds (%rax){1to8}, %ymm1, %ymm2 {k1} +vpdpwssds %ymm0, %ymm1, %ymm2 {z}{k1} +vpdpwssds (%rax), %ymm1, %ymm2 {z}{k1} +vpdpwssds (%rax){1to8}, %ymm1, %ymm2 {z}{k1} + +# CHECK: Instruction Info: +# CHECK-NEXT: [1]: #uOps +# CHECK-NEXT: [2]: Latency +# CHECK-NEXT: [3]: RThroughput +# CHECK-NEXT: [4]: MayLoad +# CHECK-NEXT: [5]: MayStore +# CHECK-NEXT: [6]: HasSideEffects (U) + +# CHECK: [1] [2] [3] [4] [5] [6] Instructions: +# CHECK-NEXT: 1 5 0.50 vpdpbusd %xmm0, %xmm1, %xmm2 +# CHECK-NEXT: 2 11 0.50 * vpdpbusd (%rax), %xmm1, %xmm2 +# CHECK-NEXT: 2 11 0.50 * vpdpbusd (%rax){1to4}, %xmm1, %xmm2 +# CHECK-NEXT: 1 5 0.50 vpdpbusd %xmm0, %xmm1, %xmm2 {%k1} +# CHECK-NEXT: 2 11 0.50 * vpdpbusd (%rax), %xmm1, %xmm2 {%k1} +# CHECK-NEXT: 2 11 0.50 * vpdpbusd (%rax){1to4}, %xmm1, %xmm2 {%k1} +# CHECK-NEXT: 1 5 0.50 vpdpbusd %xmm0, %xmm1, %xmm2 {%k1} {z} +# CHECK-NEXT: 2 11 0.50 * vpdpbusd (%rax), %xmm1, %xmm2 {%k1} {z} +# CHECK-NEXT: 2 11 0.50 * vpdpbusd (%rax){1to4}, %xmm1, %xmm2 {%k1} {z} +# CHECK-NEXT: 1 5 0.50 vpdpbusd %ymm0, %ymm1, %ymm2 +# CHECK-NEXT: 2 12 0.50 * vpdpbusd (%rax), %ymm1, %ymm2 +# CHECK-NEXT: 2 12 0.50 * vpdpbusd (%rax){1to8}, %ymm1, %ymm2 +# CHECK-NEXT: 1 5 0.50 vpdpbusd %ymm0, %ymm1, %ymm2 {%k1} +# CHECK-NEXT: 2 12 0.50 * vpdpbusd (%rax), %ymm1, %ymm2 {%k1} +# CHECK-NEXT: 2 12 0.50 * vpdpbusd (%rax){1to8}, %ymm1, %ymm2 {%k1} +# CHECK-NEXT: 1 5 0.50 vpdpbusd %ymm0, %ymm1, %ymm2 {%k1} {z} +# CHECK-NEXT: 2 12 0.50 * vpdpbusd (%rax), %ymm1, %ymm2 {%k1} {z} +# CHECK-NEXT: 2 12 0.50 * vpdpbusd (%rax){1to8}, %ymm1, %ymm2 {%k1} {z} +# CHECK-NEXT: 1 5 0.50 vpdpbusds %xmm0, %xmm1, %xmm2 +# CHECK-NEXT: 2 11 0.50 * vpdpbusds (%rax), %xmm1, %xmm2 +# CHECK-NEXT: 2 11 0.50 * vpdpbusds (%rax){1to4}, %xmm1, %xmm2 +# CHECK-NEXT: 1 5 0.50 vpdpbusds %xmm0, %xmm1, %xmm2 {%k1} +# CHECK-NEXT: 2 11 0.50 * vpdpbusds (%rax), %xmm1, %xmm2 {%k1} +# CHECK-NEXT: 2 11 0.50 * vpdpbusds (%rax){1to4}, %xmm1, %xmm2 {%k1} +# CHECK-NEXT: 1 5 0.50 vpdpbusds %xmm0, %xmm1, %xmm2 {%k1} {z} +# CHECK-NEXT: 2 11 0.50 * vpdpbusds (%rax), %xmm1, %xmm2 {%k1} {z} +# CHECK-NEXT: 2 11 0.50 * vpdpbusds (%rax){1to4}, %xmm1, %xmm2 {%k1} {z} +# CHECK-NEXT: 1 5 0.50 vpdpbusds %ymm0, %ymm1, %ymm2 +# CHECK-NEXT: 2 12 0.50 * vpdpbusds (%rax), %ymm1, %ymm2 +# CHECK-NEXT: 2 12 0.50 * vpdpbusds (%rax){1to8}, %ymm1, %ymm2 +# CHECK-NEXT: 1 5 0.50 vpdpbusds %ymm0, %ymm1, %ymm2 {%k1} +# CHECK-NEXT: 2 12 0.50 * vpdpbusds (%rax), %ymm1, %ymm2 {%k1} +# CHECK-NEXT: 2 12 0.50 * vpdpbusds (%rax){1to8}, %ymm1, %ymm2 {%k1} +# CHECK-NEXT: 1 5 0.50 vpdpbusds %ymm0, %ymm1, %ymm2 {%k1} {z} +# CHECK-NEXT: 2 12 0.50 * vpdpbusds (%rax), %ymm1, %ymm2 {%k1} {z} +# CHECK-NEXT: 2 12 0.50 * vpdpbusds (%rax){1to8}, %ymm1, %ymm2 {%k1} {z} +# CHECK-NEXT: 1 5 0.50 vpdpwssd %xmm0, %xmm1, %xmm2 +# CHECK-NEXT: 2 11 0.50 * vpdpwssd (%rax), %xmm1, %xmm2 +# CHECK-NEXT: 2 11 0.50 * vpdpwssd (%rax){1to4}, %xmm1, %xmm2 +# CHECK-NEXT: 1 5 0.50 vpdpwssd %xmm0, %xmm1, %xmm2 {%k1} +# CHECK-NEXT: 2 11 0.50 * vpdpwssd (%rax), %xmm1, %xmm2 {%k1} +# CHECK-NEXT: 2 11 0.50 * vpdpwssd (%rax){1to4}, %xmm1, %xmm2 {%k1} +# CHECK-NEXT: 1 5 0.50 vpdpwssd %xmm0, %xmm1, %xmm2 {%k1} {z} +# CHECK-NEXT: 2 11 0.50 * vpdpwssd (%rax), %xmm1, %xmm2 {%k1} {z} +# CHECK-NEXT: 2 11 0.50 * vpdpwssd (%rax){1to4}, %xmm1, %xmm2 {%k1} {z} +# CHECK-NEXT: 1 5 0.50 vpdpwssd %ymm0, %ymm1, %ymm2 +# CHECK-NEXT: 2 12 0.50 * vpdpwssd (%rax), %ymm1, %ymm2 +# CHECK-NEXT: 2 12 0.50 * vpdpwssd (%rax){1to8}, %ymm1, %ymm2 +# CHECK-NEXT: 1 5 0.50 vpdpwssd %ymm0, %ymm1, %ymm2 {%k1} +# CHECK-NEXT: 2 12 0.50 * vpdpwssd (%rax), %ymm1, %ymm2 {%k1} +# CHECK-NEXT: 2 12 0.50 * vpdpwssd (%rax){1to8}, %ymm1, %ymm2 {%k1} +# CHECK-NEXT: 1 5 0.50 vpdpwssd %ymm0, %ymm1, %ymm2 {%k1} {z} +# CHECK-NEXT: 2 12 0.50 * vpdpwssd (%rax), %ymm1, %ymm2 {%k1} {z} +# CHECK-NEXT: 2 12 0.50 * vpdpwssd (%rax){1to8}, %ymm1, %ymm2 {%k1} {z} +# CHECK-NEXT: 1 5 0.50 vpdpwssds %xmm0, %xmm1, %xmm2 +# CHECK-NEXT: 2 11 0.50 * vpdpwssds (%rax), %xmm1, %xmm2 +# CHECK-NEXT: 2 11 0.50 * vpdpwssds (%rax){1to4}, %xmm1, %xmm2 +# CHECK-NEXT: 1 5 0.50 vpdpwssds %xmm0, %xmm1, %xmm2 {%k1} +# CHECK-NEXT: 2 11 0.50 * vpdpwssds (%rax), %xmm1, %xmm2 {%k1} +# CHECK-NEXT: 2 11 0.50 * vpdpwssds (%rax){1to4}, %xmm1, %xmm2 {%k1} +# CHECK-NEXT: 1 5 0.50 vpdpwssds %xmm0, %xmm1, %xmm2 {%k1} {z} +# CHECK-NEXT: 2 11 0.50 * vpdpwssds (%rax), %xmm1, %xmm2 {%k1} {z} +# CHECK-NEXT: 2 11 0.50 * vpdpwssds (%rax){1to4}, %xmm1, %xmm2 {%k1} {z} +# CHECK-NEXT: 1 5 0.50 vpdpwssds %ymm0, %ymm1, %ymm2 +# CHECK-NEXT: 2 12 0.50 * vpdpwssds (%rax), %ymm1, %ymm2 +# CHECK-NEXT: 2 12 0.50 * vpdpwssds (%rax){1to8}, %ymm1, %ymm2 +# CHECK-NEXT: 1 5 0.50 vpdpwssds %ymm0, %ymm1, %ymm2 {%k1} +# CHECK-NEXT: 2 12 0.50 * vpdpwssds (%rax), %ymm1, %ymm2 {%k1} +# CHECK-NEXT: 2 12 0.50 * vpdpwssds (%rax){1to8}, %ymm1, %ymm2 {%k1} +# CHECK-NEXT: 1 5 0.50 vpdpwssds %ymm0, %ymm1, %ymm2 {%k1} {z} +# CHECK-NEXT: 2 12 0.50 * vpdpwssds (%rax), %ymm1, %ymm2 {%k1} {z} +# CHECK-NEXT: 2 12 0.50 * vpdpwssds (%rax){1to8}, %ymm1, %ymm2 {%k1} {z} + +# CHECK: Resources: +# CHECK-NEXT: [0] - ICXDivider +# CHECK-NEXT: [1] - ICXFPDivider +# CHECK-NEXT: [2] - ICXPort0 +# CHECK-NEXT: [3] - ICXPort1 +# CHECK-NEXT: [4] - ICXPort2 +# CHECK-NEXT: [5] - ICXPort3 +# CHECK-NEXT: [6] - ICXPort4 +# CHECK-NEXT: [7] - ICXPort5 +# CHECK-NEXT: [8] - ICXPort6 +# CHECK-NEXT: [9] - ICXPort7 +# CHECK-NEXT: [10] - ICXPort8 +# CHECK-NEXT: [11] - ICXPort9 + +# CHECK: Resource pressure per iteration: +# CHECK-NEXT: [0] [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] +# CHECK-NEXT: - - 36.00 36.00 24.00 24.00 - - - - - - + +# CHECK: Resource pressure by instruction: +# CHECK-NEXT: [0] [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] Instructions: +# CHECK-NEXT: - - 0.50 0.50 - - - - - - - - vpdpbusd %xmm0, %xmm1, %xmm2 +# CHECK-NEXT: - - 0.50 0.50 0.50 0.50 - - - - - - vpdpbusd (%rax), %xmm1, %xmm2 +# CHECK-NEXT: - - 0.50 0.50 0.50 0.50 - - - - - - vpdpbusd (%rax){1to4}, %xmm1, %xmm2 +# CHECK-NEXT: - - 0.50 0.50 - - - - - - - - vpdpbusd %xmm0, %xmm1, %xmm2 {%k1} +# CHECK-NEXT: - - 0.50 0.50 0.50 0.50 - - - - - - vpdpbusd (%rax), %xmm1, %xmm2 {%k1} +# CHECK-NEXT: - - 0.50 0.50 0.50 0.50 - - - - - - vpdpbusd (%rax){1to4}, %xmm1, %xmm2 {%k1} +# CHECK-NEXT: - - 0.50 0.50 - - - - - - - - vpdpbusd %xmm0, %xmm1, %xmm2 {%k1} {z} +# CHECK-NEXT: - - 0.50 0.50 0.50 0.50 - - - - - - vpdpbusd (%rax), %xmm1, %xmm2 {%k1} {z} +# CHECK-NEXT: - - 0.50 0.50 0.50 0.50 - - - - - - vpdpbusd (%rax){1to4}, %xmm1, %xmm2 {%k1} {z} +# CHECK-NEXT: - - 0.50 0.50 - - - - - - - - vpdpbusd %ymm0, %ymm1, %ymm2 +# CHECK-NEXT: - - 0.50 0.50 0.50 0.50 - - - - - - vpdpbusd (%rax), %ymm1, %ymm2 +# CHECK-NEXT: - - 0.50 0.50 0.50 0.50 - - - - - - vpdpbusd (%rax){1to8}, %ymm1, %ymm2 +# CHECK-NEXT: - - 0.50 0.50 - - - - - - - - vpdpbusd %ymm0, %ymm1, %ymm2 {%k1} +# CHECK-NEXT: - - 0.50 0.50 0.50 0.50 - - - - - - vpdpbusd (%rax), %ymm1, %ymm2 {%k1} +# CHECK-NEXT: - - 0.50 0.50 0.50 0.50 - - - - - - vpdpbusd (%rax){1to8}, %ymm1, %ymm2 {%k1} +# CHECK-NEXT: - - 0.50 0.50 - - - - - - - - vpdpbusd %ymm0, %ymm1, %ymm2 {%k1} {z} +# CHECK-NEXT: - - 0.50 0.50 0.50 0.50 - - - - - - vpdpbusd (%rax), %ymm1, %ymm2 {%k1} {z} +# CHECK-NEXT: - - 0.50 0.50 0.50 0.50 - - - - - - vpdpbusd (%rax){1to8}, %ymm1, %ymm2 {%k1} {z} +# CHECK-NEXT: - - 0.50 0.50 - - - - - - - - vpdpbusds %xmm0, %xmm1, %xmm2 +# CHECK-NEXT: - - 0.50 0.50 0.50 0.50 - - - - - - vpdpbusds (%rax), %xmm1, %xmm2 +# CHECK-NEXT: - - 0.50 0.50 0.50 0.50 - - - - - - vpdpbusds (%rax){1to4}, %xmm1, %xmm2 +# CHECK-NEXT: - - 0.50 0.50 - - - - - - - - vpdpbusds %xmm0, %xmm1, %xmm2 {%k1} +# CHECK-NEXT: - - 0.50 0.50 0.50 0.50 - - - - - - vpdpbusds (%rax), %xmm1, %xmm2 {%k1} +# CHECK-NEXT: - - 0.50 0.50 0.50 0.50 - - - - - - vpdpbusds (%rax){1to4}, %xmm1, %xmm2 {%k1} +# CHECK-NEXT: - - 0.50 0.50 - - - - - - - - vpdpbusds %xmm0, %xmm1, %xmm2 {%k1} {z} +# CHECK-NEXT: - - 0.50 0.50 0.50 0.50 - - - - - - vpdpbusds (%rax), %xmm1, %xmm2 {%k1} {z} +# CHECK-NEXT: - - 0.50 0.50 0.50 0.50 - - - - - - vpdpbusds (%rax){1to4}, %xmm1, %xmm2 {%k1} {z} +# CHECK-NEXT: - - 0.50 0.50 - - - - - - - - vpdpbusds %ymm0, %ymm1, %ymm2 +# CHECK-NEXT: - - 0.50 0.50 0.50 0.50 - - - - - - vpdpbusds (%rax), %ymm1, %ymm2 +# CHECK-NEXT: - - 0.50 0.50 0.50 0.50 - - - - - - vpdpbusds (%rax){1to8}, %ymm1, %ymm2 +# CHECK-NEXT: - - 0.50 0.50 - - - - - - - - vpdpbusds %ymm0, %ymm1, %ymm2 {%k1} +# CHECK-NEXT: - - 0.50 0.50 0.50 0.50 - - - - - - vpdpbusds (%rax), %ymm1, %ymm2 {%k1} +# CHECK-NEXT: - - 0.50 0.50 0.50 0.50 - - - - - - vpdpbusds (%rax){1to8}, %ymm1, %ymm2 {%k1} +# CHECK-NEXT: - - 0.50 0.50 - - - - - - - - vpdpbusds %ymm0, %ymm1, %ymm2 {%k1} {z} +# CHECK-NEXT: - - 0.50 0.50 0.50 0.50 - - - - - - vpdpbusds (%rax), %ymm1, %ymm2 {%k1} {z} +# CHECK-NEXT: - - 0.50 0.50 0.50 0.50 - - - - - - vpdpbusds (%rax){1to8}, %ymm1, %ymm2 {%k1} {z} +# CHECK-NEXT: - - 0.50 0.50 - - - - - - - - vpdpwssd %xmm0, %xmm1, %xmm2 +# CHECK-NEXT: - - 0.50 0.50 0.50 0.50 - - - - - - vpdpwssd (%rax), %xmm1, %xmm2 +# CHECK-NEXT: - - 0.50 0.50 0.50 0.50 - - - - - - vpdpwssd (%rax){1to4}, %xmm1, %xmm2 +# CHECK-NEXT: - - 0.50 0.50 - - - - - - - - vpdpwssd %xmm0, %xmm1, %xmm2 {%k1} +# CHECK-NEXT: - - 0.50 0.50 0.50 0.50 - - - - - - vpdpwssd (%rax), %xmm1, %xmm2 {%k1} +# CHECK-NEXT: - - 0.50 0.50 0.50 0.50 - - - - - - vpdpwssd (%rax){1to4}, %xmm1, %xmm2 {%k1} +# CHECK-NEXT: - - 0.50 0.50 - - - - - - - - vpdpwssd %xmm0, %xmm1, %xmm2 {%k1} {z} +# CHECK-NEXT: - - 0.50 0.50 0.50 0.50 - - - - - - vpdpwssd (%rax), %xmm1, %xmm2 {%k1} {z} +# CHECK-NEXT: - - 0.50 0.50 0.50 0.50 - - - - - - vpdpwssd (%rax){1to4}, %xmm1, %xmm2 {%k1} {z} +# CHECK-NEXT: - - 0.50 0.50 - - - - - - - - vpdpwssd %ymm0, %ymm1, %ymm2 +# CHECK-NEXT: - - 0.50 0.50 0.50 0.50 - - - - - - vpdpwssd (%rax), %ymm1, %ymm2 +# CHECK-NEXT: - - 0.50 0.50 0.50 0.50 - - - - - - vpdpwssd (%rax){1to8}, %ymm1, %ymm2 +# CHECK-NEXT: - - 0.50 0.50 - - - - - - - - vpdpwssd %ymm0, %ymm1, %ymm2 {%k1} +# CHECK-NEXT: - - 0.50 0.50 0.50 0.50 - - - - - - vpdpwssd (%rax), %ymm1, %ymm2 {%k1} +# CHECK-NEXT: - - 0.50 0.50 0.50 0.50 - - - - - - vpdpwssd (%rax){1to8}, %ymm1, %ymm2 {%k1} +# CHECK-NEXT: - - 0.50 0.50 - - - - - - - - vpdpwssd %ymm0, %ymm1, %ymm2 {%k1} {z} +# CHECK-NEXT: - - 0.50 0.50 0.50 0.50 - - - - - - vpdpwssd (%rax), %ymm1, %ymm2 {%k1} {z} +# CHECK-NEXT: - - 0.50 0.50 0.50 0.50 - - - - - - vpdpwssd (%rax){1to8}, %ymm1, %ymm2 {%k1} {z} +# CHECK-NEXT: - - 0.50 0.50 - - - - - - - - vpdpwssds %xmm0, %xmm1, %xmm2 +# CHECK-NEXT: - - 0.50 0.50 0.50 0.50 - - - - - - vpdpwssds (%rax), %xmm1, %xmm2 +# CHECK-NEXT: - - 0.50 0.50 0.50 0.50 - - - - - - vpdpwssds (%rax){1to4}, %xmm1, %xmm2 +# CHECK-NEXT: - - 0.50 0.50 - - - - - - - - vpdpwssds %xmm0, %xmm1, %xmm2 {%k1} +# CHECK-NEXT: - - 0.50 0.50 0.50 0.50 - - - - - - vpdpwssds (%rax), %xmm1, %xmm2 {%k1} +# CHECK-NEXT: - - 0.50 0.50 0.50 0.50 - - - - - - vpdpwssds (%rax){1to4}, %xmm1, %xmm2 {%k1} +# CHECK-NEXT: - - 0.50 0.50 - - - - - - - - vpdpwssds %xmm0, %xmm1, %xmm2 {%k1} {z} +# CHECK-NEXT: - - 0.50 0.50 0.50 0.50 - - - - - - vpdpwssds (%rax), %xmm1, %xmm2 {%k1} {z} +# CHECK-NEXT: - - 0.50 0.50 0.50 0.50 - - - - - - vpdpwssds (%rax){1to4}, %xmm1, %xmm2 {%k1} {z} +# CHECK-NEXT: - - 0.50 0.50 - - - - - - - - vpdpwssds %ymm0, %ymm1, %ymm2 +# CHECK-NEXT: - - 0.50 0.50 0.50 0.50 - - - - - - vpdpwssds (%rax), %ymm1, %ymm2 +# CHECK-NEXT: - - 0.50 0.50 0.50 0.50 - - - - - - vpdpwssds (%rax){1to8}, %ymm1, %ymm2 +# CHECK-NEXT: - - 0.50 0.50 - - - - - - - - vpdpwssds %ymm0, %ymm1, %ymm2 {%k1} +# CHECK-NEXT: - - 0.50 0.50 0.50 0.50 - - - - - - vpdpwssds (%rax), %ymm1, %ymm2 {%k1} +# CHECK-NEXT: - - 0.50 0.50 0.50 0.50 - - - - - - vpdpwssds (%rax){1to8}, %ymm1, %ymm2 {%k1} +# CHECK-NEXT: - - 0.50 0.50 - - - - - - - - vpdpwssds %ymm0, %ymm1, %ymm2 {%k1} {z} +# CHECK-NEXT: - - 0.50 0.50 0.50 0.50 - - - - - - vpdpwssds (%rax), %ymm1, %ymm2 {%k1} {z} +# CHECK-NEXT: - - 0.50 0.50 0.50 0.50 - - - - - - vpdpwssds (%rax){1to8}, %ymm1, %ymm2 {%k1} {z} diff --git a/llvm/test/tools/llvm-mca/X86/IceLakeServer/resources-avxvnni.s b/llvm/test/tools/llvm-mca/X86/IceLakeServer/resources-avxvnni.s new file mode 100644 index 000000000000..1baac8837a41 --- /dev/null +++ b/llvm/test/tools/llvm-mca/X86/IceLakeServer/resources-avxvnni.s @@ -0,0 +1,89 @@ +# NOTE: Assertions have been autogenerated by utils/update_mca_test_checks.py +# RUN: llvm-mca -mtriple=x86_64-unknown-unknown -mcpu=icelake-server -instruction-tables < %s | FileCheck %s + +vpdpbusd %xmm0, %xmm1, %xmm2 +vpdpbusd (%rax), %xmm1, %xmm2 + +vpdpbusd %ymm0, %ymm1, %ymm2 +vpdpbusd (%rax), %ymm1, %ymm2 + +vpdpbusds %xmm0, %xmm1, %xmm2 +vpdpbusds (%rax), %xmm1, %xmm2 + +vpdpbusds %ymm0, %ymm1, %ymm2 +vpdpbusds (%rax), %ymm1, %ymm2 + +vpdpwssd %xmm0, %xmm1, %xmm2 +vpdpwssd (%rax), %xmm1, %xmm2 + +vpdpwssd %ymm0, %ymm1, %ymm2 +vpdpwssd (%rax), %ymm1, %ymm2 + +vpdpwssds %xmm0, %xmm1, %xmm2 +vpdpwssds (%rax), %xmm1, %xmm2 + +vpdpwssds %ymm0, %ymm1, %ymm2 +vpdpwssds (%rax), %ymm1, %ymm2 + +# CHECK: Instruction Info: +# CHECK-NEXT: [1]: #uOps +# CHECK-NEXT: [2]: Latency +# CHECK-NEXT: [3]: RThroughput +# CHECK-NEXT: [4]: MayLoad +# CHECK-NEXT: [5]: MayStore +# CHECK-NEXT: [6]: HasSideEffects (U) + +# CHECK: [1] [2] [3] [4] [5] [6] Instructions: +# CHECK-NEXT: 1 5 0.50 vpdpbusd %xmm0, %xmm1, %xmm2 +# CHECK-NEXT: 2 11 0.50 * vpdpbusd (%rax), %xmm1, %xmm2 +# CHECK-NEXT: 1 5 0.50 vpdpbusd %ymm0, %ymm1, %ymm2 +# CHECK-NEXT: 2 12 0.50 * vpdpbusd (%rax), %ymm1, %ymm2 +# CHECK-NEXT: 1 5 0.50 vpdpbusds %xmm0, %xmm1, %xmm2 +# CHECK-NEXT: 2 11 0.50 * vpdpbusds (%rax), %xmm1, %xmm2 +# CHECK-NEXT: 1 5 0.50 vpdpbusds %ymm0, %ymm1, %ymm2 +# CHECK-NEXT: 2 12 0.50 * vpdpbusds (%rax), %ymm1, %ymm2 +# CHECK-NEXT: 1 5 0.50 vpdpwssd %xmm0, %xmm1, %xmm2 +# CHECK-NEXT: 2 11 0.50 * vpdpwssd (%rax), %xmm1, %xmm2 +# CHECK-NEXT: 1 5 0.50 vpdpwssd %ymm0, %ymm1, %ymm2 +# CHECK-NEXT: 2 12 0.50 * vpdpwssd (%rax), %ymm1, %ymm2 +# CHECK-NEXT: 1 5 0.50 vpdpwssds %xmm0, %xmm1, %xmm2 +# CHECK-NEXT: 2 11 0.50 * vpdpwssds (%rax), %xmm1, %xmm2 +# CHECK-NEXT: 1 5 0.50 vpdpwssds %ymm0, %ymm1, %ymm2 +# CHECK-NEXT: 2 12 0.50 * vpdpwssds (%rax), %ymm1, %ymm2 + +# CHECK: Resources: +# CHECK-NEXT: [0] - ICXDivider +# CHECK-NEXT: [1] - ICXFPDivider +# CHECK-NEXT: [2] - ICXPort0 +# CHECK-NEXT: [3] - ICXPort1 +# CHECK-NEXT: [4] - ICXPort2 +# CHECK-NEXT: [5] - ICXPort3 +# CHECK-NEXT: [6] - ICXPort4 +# CHECK-NEXT: [7] - ICXPort5 +# CHECK-NEXT: [8] - ICXPort6 +# CHECK-NEXT: [9] - ICXPort7 +# CHECK-NEXT: [10] - ICXPort8 +# CHECK-NEXT: [11] - ICXPort9 + +# CHECK: Resource pressure per iteration: +# CHECK-NEXT: [0] [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] +# CHECK-NEXT: - - 8.00 8.00 4.00 4.00 - - - - - - + +# CHECK: Resource pressure by instruction: +# CHECK-NEXT: [0] [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] Instructions: +# CHECK-NEXT: - - 0.50 0.50 - - - - - - - - vpdpbusd %xmm0, %xmm1, %xmm2 +# CHECK-NEXT: - - 0.50 0.50 0.50 0.50 - - - - - - vpdpbusd (%rax), %xmm1, %xmm2 +# CHECK-NEXT: - - 0.50 0.50 - - - - - - - - vpdpbusd %ymm0, %ymm1, %ymm2 +# CHECK-NEXT: - - 0.50 0.50 0.50 0.50 - - - - - - vpdpbusd (%rax), %ymm1, %ymm2 +# CHECK-NEXT: - - 0.50 0.50 - - - - - - - - vpdpbusds %xmm0, %xmm1, %xmm2 +# CHECK-NEXT: - - 0.50 0.50 0.50 0.50 - - - - - - vpdpbusds (%rax), %xmm1, %xmm2 +# CHECK-NEXT: - - 0.50 0.50 - - - - - - - - vpdpbusds %ymm0, %ymm1, %ymm2 +# CHECK-NEXT: - - 0.50 0.50 0.50 0.50 - - - - - - vpdpbusds (%rax), %ymm1, %ymm2 +# CHECK-NEXT: - - 0.50 0.50 - - - - - - - - vpdpwssd %xmm0, %xmm1, %xmm2 +# CHECK-NEXT: - - 0.50 0.50 0.50 0.50 - - - - - - vpdpwssd (%rax), %xmm1, %xmm2 +# CHECK-NEXT: - - 0.50 0.50 - - - - - - - - vpdpwssd %ymm0, %ymm1, %ymm2 +# CHECK-NEXT: - - 0.50 0.50 0.50 0.50 - - - - - - vpdpwssd (%rax), %ymm1, %ymm2 +# CHECK-NEXT: - - 0.50 0.50 - - - - - - - - vpdpwssds %xmm0, %xmm1, %xmm2 +# CHECK-NEXT: - - 0.50 0.50 0.50 0.50 - - - - - - vpdpwssds (%rax), %xmm1, %xmm2 +# CHECK-NEXT: - - 0.50 0.50 - - - - - - - - vpdpwssds %ymm0, %ymm1, %ymm2 +# CHECK-NEXT: - - 0.50 0.50 0.50 0.50 - - - - - - vpdpwssds (%rax), %ymm1, %ymm2 -- cgit v1.2.3