mirror of
https://github.com/RPCS3/llvm-mirror.git
synced 2024-11-22 18:54:02 +01:00
a3da92d4e3
This patch updates X86InstCombineIntrinsic.cpp to use the newly updated CreateShuffleVector. The tests are updated because the updated CreateShuffleVector uses poison value for the second vector. If I didn't miss something, the masks in the tests are choosing elements from the first vector only; therefore the tests are having equivalent behavior. Reviewed By: RKSimon Differential Revision: https://reviews.llvm.org/D94059
111 lines
5.0 KiB
LLVM
111 lines
5.0 KiB
LLVM
; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
|
|
; RUN: opt < %s -instcombine -mtriple=x86_64-unknown-unknown -S | FileCheck %s
|
|
target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
|
|
|
|
; Verify that instcombine is able to fold identity shuffles.
|
|
|
|
define <8 x i32> @identity_test_vpermd(<8 x i32> %a0) {
|
|
; CHECK-LABEL: @identity_test_vpermd(
|
|
; CHECK-NEXT: ret <8 x i32> [[A0:%.*]]
|
|
;
|
|
%a = tail call <8 x i32> @llvm.x86.avx2.permd(<8 x i32> %a0, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>)
|
|
ret <8 x i32> %a
|
|
}
|
|
|
|
define <8 x float> @identity_test_vpermps(<8 x float> %a0) {
|
|
; CHECK-LABEL: @identity_test_vpermps(
|
|
; CHECK-NEXT: ret <8 x float> [[A0:%.*]]
|
|
;
|
|
%a = tail call <8 x float> @llvm.x86.avx2.permps(<8 x float> %a0, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>)
|
|
ret <8 x float> %a
|
|
}
|
|
|
|
; Instcombine should be able to fold the following shuffle to a builtin shufflevector
|
|
; with a shuffle mask of all zeroes.
|
|
|
|
define <8 x i32> @zero_test_vpermd(<8 x i32> %a0) {
|
|
; CHECK-LABEL: @zero_test_vpermd(
|
|
; CHECK-NEXT: [[TMP1:%.*]] = shufflevector <8 x i32> [[A0:%.*]], <8 x i32> poison, <8 x i32> zeroinitializer
|
|
; CHECK-NEXT: ret <8 x i32> [[TMP1]]
|
|
;
|
|
%a = tail call <8 x i32> @llvm.x86.avx2.permd(<8 x i32> %a0, <8 x i32> zeroinitializer)
|
|
ret <8 x i32> %a
|
|
}
|
|
|
|
define <8 x float> @zero_test_vpermps(<8 x float> %a0) {
|
|
; CHECK-LABEL: @zero_test_vpermps(
|
|
; CHECK-NEXT: [[TMP1:%.*]] = shufflevector <8 x float> [[A0:%.*]], <8 x float> poison, <8 x i32> zeroinitializer
|
|
; CHECK-NEXT: ret <8 x float> [[TMP1]]
|
|
;
|
|
%a = tail call <8 x float> @llvm.x86.avx2.permps(<8 x float> %a0, <8 x i32> zeroinitializer)
|
|
ret <8 x float> %a
|
|
}
|
|
|
|
; Verify that instcombine is able to fold constant shuffles.
|
|
|
|
define <8 x i32> @shuffle_test_vpermd(<8 x i32> %a0) {
|
|
; CHECK-LABEL: @shuffle_test_vpermd(
|
|
; CHECK-NEXT: [[TMP1:%.*]] = shufflevector <8 x i32> [[A0:%.*]], <8 x i32> poison, <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
|
|
; CHECK-NEXT: ret <8 x i32> [[TMP1]]
|
|
;
|
|
%a = tail call <8 x i32> @llvm.x86.avx2.permd(<8 x i32> %a0, <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>)
|
|
ret <8 x i32> %a
|
|
}
|
|
|
|
define <8 x float> @shuffle_test_vpermps(<8 x float> %a0) {
|
|
; CHECK-LABEL: @shuffle_test_vpermps(
|
|
; CHECK-NEXT: [[TMP1:%.*]] = shufflevector <8 x float> [[A0:%.*]], <8 x float> poison, <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
|
|
; CHECK-NEXT: ret <8 x float> [[TMP1]]
|
|
;
|
|
%a = tail call <8 x float> @llvm.x86.avx2.permps(<8 x float> %a0, <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>)
|
|
ret <8 x float> %a
|
|
}
|
|
|
|
; Verify that instcombine is able to fold constant shuffles with undef mask elements.
|
|
|
|
define <8 x i32> @undef_test_vpermd(<8 x i32> %a0) {
|
|
; CHECK-LABEL: @undef_test_vpermd(
|
|
; CHECK-NEXT: [[TMP1:%.*]] = shufflevector <8 x i32> [[A0:%.*]], <8 x i32> poison, <8 x i32> <i32 undef, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
|
|
; CHECK-NEXT: ret <8 x i32> [[TMP1]]
|
|
;
|
|
%a = tail call <8 x i32> @llvm.x86.avx2.permd(<8 x i32> %a0, <8 x i32> <i32 undef, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>)
|
|
ret <8 x i32> %a
|
|
}
|
|
|
|
define <8 x float> @undef_test_vpermps(<8 x float> %a0) {
|
|
; CHECK-LABEL: @undef_test_vpermps(
|
|
; CHECK-NEXT: [[TMP1:%.*]] = shufflevector <8 x float> [[A0:%.*]], <8 x float> poison, <8 x i32> <i32 undef, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
|
|
; CHECK-NEXT: ret <8 x float> [[TMP1]]
|
|
;
|
|
%a = tail call <8 x float> @llvm.x86.avx2.permps(<8 x float> %a0, <8 x i32> <i32 undef, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>)
|
|
ret <8 x float> %a
|
|
}
|
|
|
|
; Verify simplify demanded elts.
|
|
|
|
define <8 x i32> @elts_test_vpermd(<8 x i32> %a0, i32 %a1) {
|
|
; CHECK-LABEL: @elts_test_vpermd(
|
|
; CHECK-NEXT: [[TMP1:%.*]] = shufflevector <8 x i32> [[A0:%.*]], <8 x i32> poison, <8 x i32> <i32 undef, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
|
|
; CHECK-NEXT: ret <8 x i32> [[TMP1]]
|
|
;
|
|
%1 = insertelement <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>, i32 %a1, i32 0
|
|
%2 = tail call <8 x i32> @llvm.x86.avx2.permd(<8 x i32> %a0, <8 x i32> %1)
|
|
%3 = shufflevector <8 x i32> %2, <8 x i32> poison, <8 x i32> <i32 undef, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
|
|
ret <8 x i32> %3
|
|
}
|
|
|
|
define <8 x float> @elts_test_vpermps(<8 x float> %a0, <8 x i32> %a1) {
|
|
; CHECK-LABEL: @elts_test_vpermps(
|
|
; CHECK-NEXT: [[TMP1:%.*]] = tail call <8 x float> @llvm.x86.avx2.permps(<8 x float> [[A0:%.*]], <8 x i32> [[A1:%.*]])
|
|
; CHECK-NEXT: [[TMP2:%.*]] = shufflevector <8 x float> [[TMP1]], <8 x float> poison, <8 x i32> zeroinitializer
|
|
; CHECK-NEXT: ret <8 x float> [[TMP2]]
|
|
;
|
|
%1 = insertelement <8 x i32> %a1, i32 0, i32 7
|
|
%2 = tail call <8 x float> @llvm.x86.avx2.permps(<8 x float> %a0, <8 x i32> %1)
|
|
%3 = shufflevector <8 x float> %2, <8 x float> poison, <8 x i32> zeroinitializer
|
|
ret <8 x float> %3
|
|
}
|
|
|
|
declare <8 x i32> @llvm.x86.avx2.permd(<8 x i32>, <8 x i32>)
|
|
declare <8 x float> @llvm.x86.avx2.permps(<8 x float>, <8 x i32>)
|