mono/linux-packaging-mono
0
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Imported Upstream version 5.18.0.167

Browse Source 
Former-commit-id: 289509151e0fee68a1b591a20c9f109c3c789d3a
This commit is contained in:
Xamarin Public Jenkins (auto-signing)
2018-10-20 08:25:10 +00:00
parent e19d552987
commit b084638f15
28489 changed files with 184 additions and 3866856 deletions
Download Patch File Download Diff File Expand all files Collapse all files

62
external/llvm/test/Transforms/SLPVectorizer/X86/PR32086.ll vendored
View File

@ -1,62 +0,0 @@
; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
; RUN: opt -slp-vectorizer -slp-vectorize-hor -slp-vectorize-hor-store -S < %s -mtriple=x86_64-unknown-linux-gnu -mcpu=bdver2 | FileCheck %s
define void @i64_simplified(i64* noalias %st, i64* noalias %ld) {
; CHECK-LABEL: @i64_simplified(
; CHECK-NEXT: [[ARRAYIDX1:%.*]] = getelementptr inbounds i64, i64* [[LD:%.*]], i64 1
; CHECK-NEXT: [[T0:%.*]] = load i64, i64* [[LD]], align 8
; CHECK-NEXT: [[T1:%.*]] = load i64, i64* [[ARRAYIDX1]], align 8
; CHECK-NEXT: [[ARRAYIDX3:%.*]] = getelementptr inbounds i64, i64* [[ST:%.*]], i64 1
; CHECK-NEXT: [[ARRAYIDX4:%.*]] = getelementptr inbounds i64, i64* [[ST]], i64 2
; CHECK-NEXT: [[ARRAYIDX5:%.*]] = getelementptr inbounds i64, i64* [[ST]], i64 3
; CHECK-NEXT: store i64 [[T0]], i64* [[ST]], align 8
; CHECK-NEXT: store i64 [[T1]], i64* [[ARRAYIDX3]], align 8
; CHECK-NEXT: store i64 [[T0]], i64* [[ARRAYIDX4]], align 8
; CHECK-NEXT: store i64 [[T1]], i64* [[ARRAYIDX5]], align 8
; CHECK-NEXT: ret void
;
%arrayidx1 = getelementptr inbounds i64, i64* %ld, i64 1
%t0 = load i64, i64* %ld, align 8
%t1 = load i64, i64* %arrayidx1, align 8
%arrayidx3 = getelementptr inbounds i64, i64* %st, i64 1
%arrayidx4 = getelementptr inbounds i64, i64* %st, i64 2
%arrayidx5 = getelementptr inbounds i64, i64* %st, i64 3
store i64 %t0, i64* %st, align 8
store i64 %t1, i64* %arrayidx3, align 8
store i64 %t0, i64* %arrayidx4, align 8
store i64 %t1, i64* %arrayidx5, align 8
ret void
}
define void @i64_simplifiedi_reversed(i64* noalias %st, i64* noalias %ld) {
; CHECK-LABEL: @i64_simplifiedi_reversed(
; CHECK-NEXT: [[ARRAYIDX1:%.*]] = getelementptr inbounds i64, i64* [[LD:%.*]], i64 1
; CHECK-NEXT: [[T0:%.*]] = load i64, i64* [[LD]], align 8
; CHECK-NEXT: [[T1:%.*]] = load i64, i64* [[ARRAYIDX1]], align 8
; CHECK-NEXT: [[ARRAYIDX3:%.*]] = getelementptr inbounds i64, i64* [[ST:%.*]], i64 1
; CHECK-NEXT: [[ARRAYIDX4:%.*]] = getelementptr inbounds i64, i64* [[ST]], i64 2
; CHECK-NEXT: [[ARRAYIDX5:%.*]] = getelementptr inbounds i64, i64* [[ST]], i64 3
; CHECK-NEXT: store i64 [[T1]], i64* [[ST]], align 8
; CHECK-NEXT: store i64 [[T0]], i64* [[ARRAYIDX3]], align 8
; CHECK-NEXT: store i64 [[T1]], i64* [[ARRAYIDX4]], align 8
; CHECK-NEXT: store i64 [[T0]], i64* [[ARRAYIDX5]], align 8
; CHECK-NEXT: ret void
;
%arrayidx1 = getelementptr inbounds i64, i64* %ld, i64 1
%t0 = load i64, i64* %ld, align 8
%t1 = load i64, i64* %arrayidx1, align 8
%arrayidx3 = getelementptr inbounds i64, i64* %st, i64 1
%arrayidx4 = getelementptr inbounds i64, i64* %st, i64 2
%arrayidx5 = getelementptr inbounds i64, i64* %st, i64 3
store i64 %t1, i64* %st, align 8
store i64 %t0, i64* %arrayidx3, align 8
store i64 %t1, i64* %arrayidx4, align 8
store i64 %t0, i64* %arrayidx5, align 8
ret void
}

98
external/llvm/test/Transforms/SLPVectorizer/X86/PR34635.ll vendored
View File

@ -1,98 +0,0 @@
; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
; RUN: opt < %s -mtriple=x86_64-unknown-linux -slp-vectorizer -S -mcpu=corei7 | FileCheck %s
define i32 @main() {
; CHECK-LABEL: @main(
; CHECK-NEXT: bb:
; CHECK-NEXT: [[TMP:%.*]] = alloca <8 x i32>, align 32
; CHECK-NEXT: [[TMP1:%.*]] = bitcast <8 x i32>* [[TMP]] to [8 x i32]*
; CHECK-NEXT: [[TMP2:%.*]] = alloca i32, align 4
; CHECK-NEXT: [[TMP3:%.*]] = bitcast <8 x i32>* [[TMP]] to i8*
; CHECK-NEXT: [[TMP4:%.*]] = getelementptr inbounds <8 x i32>, <8 x i32>* [[TMP]], i64 0, i64 0
; CHECK-NEXT: [[TMP5:%.*]] = getelementptr inbounds [8 x i32], [8 x i32]* [[TMP1]], i64 0, i64 1
; CHECK-NEXT: [[TMP6:%.*]] = getelementptr inbounds [8 x i32], [8 x i32]* [[TMP1]], i64 0, i64 2
; CHECK-NEXT: [[TMP7:%.*]] = getelementptr inbounds [8 x i32], [8 x i32]* [[TMP1]], i64 0, i64 3
; CHECK-NEXT: [[TMP8:%.*]] = getelementptr inbounds [8 x i32], [8 x i32]* [[TMP1]], i64 0, i64 4
; CHECK-NEXT: [[TMP9:%.*]] = getelementptr inbounds [8 x i32], [8 x i32]* [[TMP1]], i64 0, i64 6
; CHECK-NEXT: [[TMP10:%.*]] = getelementptr inbounds [8 x i32], [8 x i32]* [[TMP1]], i64 0, i64 5
; CHECK-NEXT: [[TMP11:%.*]] = getelementptr inbounds [8 x i32], [8 x i32]* [[TMP1]], i64 0, i64 7
; CHECK-NEXT: store <8 x i32> <i32 -221320154, i32 -756426931, i32 563883532, i32 382683935, i32 144890241, i32 -1052877364, i32 -1052877364, i32 -1016007675>, <8 x i32>* [[TMP]], align 32
; CHECK-NEXT: [[TMP12:%.*]] = bitcast i32* [[TMP2]] to i8*
; CHECK-NEXT: [[TMP13:%.*]] = load i32, i32* [[TMP4]], align 32
; CHECK-NEXT: [[TMP14:%.*]] = load i32, i32* [[TMP5]], align 4
; CHECK-NEXT: [[TMP15:%.*]] = icmp slt i32 [[TMP14]], [[TMP13]]
; CHECK-NEXT: [[TMP16:%.*]] = select i1 [[TMP15]], i32 [[TMP14]], i32 [[TMP13]]
; CHECK-NEXT: [[TMP17:%.*]] = zext i1 [[TMP15]] to i32
; CHECK-NEXT: [[TMP18:%.*]] = load i32, i32* [[TMP6]], align 8
; CHECK-NEXT: [[TMP19:%.*]] = icmp slt i32 [[TMP18]], [[TMP16]]
; CHECK-NEXT: [[TMP20:%.*]] = select i1 [[TMP19]], i32 [[TMP18]], i32 [[TMP16]]
; CHECK-NEXT: [[TMP21:%.*]] = select i1 [[TMP19]], i32 2, i32 [[TMP16]]
; CHECK-NEXT: [[TMP22:%.*]] = load i32, i32* [[TMP7]], align 4
; CHECK-NEXT: [[TMP23:%.*]] = icmp slt i32 [[TMP22]], [[TMP20]]
; CHECK-NEXT: [[TMP24:%.*]] = select i1 [[TMP23]], i32 [[TMP22]], i32 [[TMP20]]
; CHECK-NEXT: [[TMP25:%.*]] = select i1 [[TMP23]], i32 3, i32 [[TMP21]]
; CHECK-NEXT: [[TMP26:%.*]] = load i32, i32* [[TMP8]], align 16
; CHECK-NEXT: [[TMP27:%.*]] = icmp slt i32 [[TMP26]], [[TMP24]]
; CHECK-NEXT: [[TMP28:%.*]] = select i1 [[TMP27]], i32 [[TMP26]], i32 [[TMP24]]
; CHECK-NEXT: [[TMP29:%.*]] = select i1 [[TMP27]], i32 4, i32 [[TMP25]]
; CHECK-NEXT: [[TMP30:%.*]] = load i32, i32* [[TMP10]], align 4
; CHECK-NEXT: [[TMP31:%.*]] = icmp slt i32 [[TMP30]], [[TMP28]]
; CHECK-NEXT: [[TMP32:%.*]] = select i1 [[TMP31]], i32 [[TMP30]], i32 [[TMP28]]
; CHECK-NEXT: [[TMP33:%.*]] = select i1 [[TMP31]], i32 5, i32 [[TMP29]]
; CHECK-NEXT: [[TMP34:%.*]] = load i32, i32* [[TMP9]], align 8
; CHECK-NEXT: [[TMP35:%.*]] = icmp slt i32 [[TMP34]], [[TMP32]]
; CHECK-NEXT: [[TMP36:%.*]] = select i1 [[TMP35]], i32 [[TMP34]], i32 [[TMP32]]
; CHECK-NEXT: [[TMP37:%.*]] = select i1 [[TMP35]], i32 6, i32 [[TMP33]]
; CHECK-NEXT: [[TMP38:%.*]] = load i32, i32* [[TMP11]], align 4
; CHECK-NEXT: [[TMP39:%.*]] = icmp slt i32 [[TMP38]], [[TMP36]]
; CHECK-NEXT: [[TMP40:%.*]] = select i1 [[TMP39]], i32 7, i32 [[TMP37]]
; CHECK-NEXT: store i32 [[TMP40]], i32* [[TMP2]], align 4
; CHECK-NEXT: ret i32 0
;
bb:
%tmp = alloca <8 x i32>, align 32
%tmp1 = bitcast <8 x i32>* %tmp to [8 x i32]*
%tmp2 = alloca i32, align 4
%tmp3 = bitcast <8 x i32>* %tmp to i8*
%tmp4 = getelementptr inbounds <8 x i32>, <8 x i32>* %tmp, i64 0, i64 0
%tmp5 = getelementptr inbounds [8 x i32], [8 x i32]* %tmp1, i64 0, i64 1
%tmp6 = getelementptr inbounds [8 x i32], [8 x i32]* %tmp1, i64 0, i64 2
%tmp7 = getelementptr inbounds [8 x i32], [8 x i32]* %tmp1, i64 0, i64 3
%tmp8 = getelementptr inbounds [8 x i32], [8 x i32]* %tmp1, i64 0, i64 4
%tmp9 = getelementptr inbounds [8 x i32], [8 x i32]* %tmp1, i64 0, i64 6
%tmp10 = getelementptr inbounds [8 x i32], [8 x i32]* %tmp1, i64 0, i64 5
%tmp11 = getelementptr inbounds [8 x i32], [8 x i32]* %tmp1, i64 0, i64 7
store <8 x i32> <i32 -221320154, i32 -756426931, i32 563883532, i32 382683935, i32 144890241, i32 -1052877364, i32 -1052877364, i32 -1016007675>, <8 x i32>* %tmp, align 32
%tmp12 = bitcast i32* %tmp2 to i8*
%tmp13 = load i32, i32* %tmp4, align 32
%tmp14 = load i32, i32* %tmp5, align 4
%tmp15 = icmp slt i32 %tmp14, %tmp13
%tmp16 = select i1 %tmp15, i32 %tmp14, i32 %tmp13
%tmp17 = zext i1 %tmp15 to i32
%tmp18 = load i32, i32* %tmp6, align 8
%tmp19 = icmp slt i32 %tmp18, %tmp16
%tmp20 = select i1 %tmp19, i32 %tmp18, i32 %tmp16
%tmp21 = select i1 %tmp19, i32 2, i32 %tmp16
%tmp22 = load i32, i32* %tmp7, align 4
%tmp23 = icmp slt i32 %tmp22, %tmp20
%tmp24 = select i1 %tmp23, i32 %tmp22, i32 %tmp20
%tmp25 = select i1 %tmp23, i32 3, i32 %tmp21
%tmp26 = load i32, i32* %tmp8, align 16
%tmp27 = icmp slt i32 %tmp26, %tmp24
%tmp28 = select i1 %tmp27, i32 %tmp26, i32 %tmp24
%tmp29 = select i1 %tmp27, i32 4, i32 %tmp25
%tmp30 = load i32, i32* %tmp10, align 4
%tmp31 = icmp slt i32 %tmp30, %tmp28
%tmp32 = select i1 %tmp31, i32 %tmp30, i32 %tmp28
%tmp33 = select i1 %tmp31, i32 5, i32 %tmp29
%tmp34 = load i32, i32* %tmp9, align 8
%tmp35 = icmp slt i32 %tmp34, %tmp32
%tmp36 = select i1 %tmp35, i32 %tmp34, i32 %tmp32
%tmp37 = select i1 %tmp35, i32 6, i32 %tmp33
%tmp38 = load i32, i32* %tmp11, align 4
%tmp39 = icmp slt i32 %tmp38, %tmp36
%tmp40 = select i1 %tmp39, i32 7, i32 %tmp37
store i32 %tmp40, i32* %tmp2, align 4
ret i32 0
}

74
external/llvm/test/Transforms/SLPVectorizer/X86/PR35628_1.ll vendored
View File

@ -1,74 +0,0 @@
; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
; RUN: opt -slp-vectorizer -slp-vectorize-hor -slp-vectorize-hor-store -S < %s -mtriple=x86_64-unknown-linux-gnu | FileCheck %s
target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128-ni:1"
define void @mainTest(i32* %ptr) #0 {
; CHECK-LABEL: @mainTest(
; CHECK-NEXT: entry:
; CHECK-NEXT: [[CMP:%.*]] = icmp eq i32* [[PTR:%.*]], null
; CHECK-NEXT: br i1 [[CMP]], label [[LOOP:%.*]], label [[BAIL_OUT:%.*]]
; CHECK: loop:
; CHECK-NEXT: [[DUMMY_PHI:%.*]] = phi i32 [ 1, [[ENTRY:%.*]] ], [ [[OP_EXTRA5:%.*]], [[LOOP]] ]
; CHECK-NEXT: [[TMP0:%.*]] = getelementptr inbounds i32, i32* [[PTR]], i64 1
; CHECK-NEXT: [[TMP1:%.*]] = getelementptr inbounds i32, i32* [[PTR]], i64 2
; CHECK-NEXT: [[TMP2:%.*]] = getelementptr inbounds i32, i32* [[PTR]], i64 3
; CHECK-NEXT: [[TMP3:%.*]] = bitcast i32* [[PTR]] to <4 x i32>*
; CHECK-NEXT: [[TMP4:%.*]] = load <4 x i32>, <4 x i32>* [[TMP3]], align 4
; CHECK-NEXT: [[TMP5:%.*]] = extractelement <4 x i32> [[TMP4]], i32 3
; CHECK-NEXT: [[TMP6:%.*]] = extractelement <4 x i32> [[TMP4]], i32 2
; CHECK-NEXT: [[TMP7:%.*]] = extractelement <4 x i32> [[TMP4]], i32 1
; CHECK-NEXT: [[TMP8:%.*]] = mul <4 x i32> [[TMP4]], [[TMP4]]
; CHECK-NEXT: [[TMP9:%.*]] = add i32 1, undef
; CHECK-NEXT: [[TMP10:%.*]] = add i32 [[TMP9]], undef
; CHECK-NEXT: [[TMP11:%.*]] = add i32 [[TMP10]], undef
; CHECK-NEXT: [[TMP12:%.*]] = add i32 [[TMP11]], undef
; CHECK-NEXT: [[TMP13:%.*]] = add i32 [[TMP12]], undef
; CHECK-NEXT: [[TMP14:%.*]] = sext i32 [[TMP6]] to i64
; CHECK-NEXT: [[TMP15:%.*]] = add i32 [[TMP13]], undef
; CHECK-NEXT: [[RDX_SHUF:%.*]] = shufflevector <4 x i32> [[TMP8]], <4 x i32> undef, <4 x i32> <i32 2, i32 3, i32 undef, i32 undef>
; CHECK-NEXT: [[BIN_RDX:%.*]] = add <4 x i32> [[TMP8]], [[RDX_SHUF]]
; CHECK-NEXT: [[RDX_SHUF1:%.*]] = shufflevector <4 x i32> [[BIN_RDX]], <4 x i32> undef, <4 x i32> <i32 1, i32 undef, i32 undef, i32 undef>
; CHECK-NEXT: [[BIN_RDX2:%.*]] = add <4 x i32> [[BIN_RDX]], [[RDX_SHUF1]]
; CHECK-NEXT: [[TMP16:%.*]] = extractelement <4 x i32> [[BIN_RDX2]], i32 0
; CHECK-NEXT: [[OP_EXTRA:%.*]] = add i32 [[TMP16]], 1
; CHECK-NEXT: [[OP_EXTRA3:%.*]] = add i32 [[OP_EXTRA]], [[TMP7]]
; CHECK-NEXT: [[OP_EXTRA4:%.*]] = add i32 [[OP_EXTRA3]], [[TMP6]]
; CHECK-NEXT: [[OP_EXTRA5]] = add i32 [[OP_EXTRA4]], [[TMP5]]
; CHECK-NEXT: [[TMP17:%.*]] = add i32 [[TMP15]], undef
; CHECK-NEXT: br label [[LOOP]]
; CHECK: bail_out:
; CHECK-NEXT: ret void
;
entry:
%cmp = icmp eq i32* %ptr, null
br i1 %cmp, label %loop, label %bail_out
loop:
%dummy_phi = phi i32 [ 1, %entry ], [ %18, %loop ]
%0 = load i32, i32 * %ptr , align 4
%1 = mul i32 %0, %0
%2 = add i32 1, %1
%3 = getelementptr inbounds i32, i32 * %ptr, i64 1
%4 = load i32, i32 * %3 , align 4
%5 = mul i32 %4, %4
%6 = add i32 %2, %4
%7 = add i32 %6, %5
%8 = getelementptr inbounds i32, i32 *%ptr, i64 2
%9 = load i32, i32 * %8 , align 4
%10 = mul i32 %9, %9
%11 = add i32 %7, %9
%12 = add i32 %11, %10
%13 = sext i32 %9 to i64
%14 = getelementptr inbounds i32, i32 *%ptr, i64 3
%15 = load i32, i32 * %14 , align 4
%16 = mul i32 %15, %15
%17 = add i32 %12, %15
%18 = add i32 %17, %16
br label %loop
bail_out:
ret void
}
attributes #0 = { "target-cpu"="westmere" }

64
external/llvm/test/Transforms/SLPVectorizer/X86/PR35628_2.ll vendored
View File

@ -1,64 +0,0 @@
; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
; RUN: opt -slp-vectorizer -slp-vectorize-hor -slp-vectorize-hor-store -S < %s -mtriple=x86_64-unknown-linux-gnu -mcpu=haswell | FileCheck %s
target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128-ni:1"
define void @test() #0 {
; CHECK-LABEL: @test(
; CHECK-NEXT: entry:
; CHECK-NEXT: br label [[LOOP:%.*]]
; CHECK: loop:
; CHECK-NEXT: [[DUMMY_PHI:%.*]] = phi i64 [ 1, [[ENTRY:%.*]] ], [ [[OP_EXTRA3:%.*]], [[LOOP]] ]
; CHECK-NEXT: [[TMP0:%.*]] = phi i64 [ 2, [[ENTRY]] ], [ [[TMP6:%.*]], [[LOOP]] ]
; CHECK-NEXT: [[DUMMY_ADD:%.*]] = add i16 0, 0
; CHECK-NEXT: [[TMP1:%.*]] = insertelement <4 x i64> undef, i64 [[TMP0]], i32 0
; CHECK-NEXT: [[TMP2:%.*]] = insertelement <4 x i64> [[TMP1]], i64 [[TMP0]], i32 1
; CHECK-NEXT: [[TMP3:%.*]] = insertelement <4 x i64> [[TMP2]], i64 [[TMP0]], i32 2
; CHECK-NEXT: [[TMP4:%.*]] = insertelement <4 x i64> [[TMP3]], i64 [[TMP0]], i32 3
; CHECK-NEXT: [[TMP5:%.*]] = add <4 x i64> <i64 3, i64 2, i64 1, i64 0>, [[TMP4]]
; CHECK-NEXT: [[TMP6]] = extractelement <4 x i64> [[TMP5]], i32 3
; CHECK-NEXT: [[TMP7:%.*]] = extractelement <4 x i64> [[TMP5]], i32 0
; CHECK-NEXT: [[DUMMY_SHL:%.*]] = shl i64 [[TMP7]], 32
; CHECK-NEXT: [[TMP8:%.*]] = add <4 x i64> <i64 1, i64 1, i64 1, i64 1>, [[TMP5]]
; CHECK-NEXT: [[TMP9:%.*]] = ashr exact <4 x i64> [[TMP8]], <i64 32, i64 32, i64 32, i64 32>
; CHECK-NEXT: [[SUM1:%.*]] = add i64 undef, undef
; CHECK-NEXT: [[SUM2:%.*]] = add i64 [[SUM1]], undef
; CHECK-NEXT: [[ZSUM:%.*]] = add i64 [[SUM2]], 0
; CHECK-NEXT: [[JOIN:%.*]] = add i64 undef, [[ZSUM]]
; CHECK-NEXT: [[RDX_SHUF:%.*]] = shufflevector <4 x i64> [[TMP9]], <4 x i64> undef, <4 x i32> <i32 2, i32 3, i32 undef, i32 undef>
; CHECK-NEXT: [[BIN_RDX:%.*]] = add <4 x i64> [[TMP9]], [[RDX_SHUF]]
; CHECK-NEXT: [[RDX_SHUF1:%.*]] = shufflevector <4 x i64> [[BIN_RDX]], <4 x i64> undef, <4 x i32> <i32 1, i32 undef, i32 undef, i32 undef>
; CHECK-NEXT: [[BIN_RDX2:%.*]] = add <4 x i64> [[BIN_RDX]], [[RDX_SHUF1]]
; CHECK-NEXT: [[TMP10:%.*]] = extractelement <4 x i64> [[BIN_RDX2]], i32 0
; CHECK-NEXT: [[OP_EXTRA:%.*]] = add i64 [[TMP10]], 0
; CHECK-NEXT: [[OP_EXTRA3]] = add i64 [[OP_EXTRA]], [[TMP6]]
; CHECK-NEXT: [[LAST:%.*]] = add i64 [[JOIN]], undef
; CHECK-NEXT: br label [[LOOP]]
;
entry:
br label %loop
loop:
%dummy_phi = phi i64 [ 1, %entry ], [ %last, %loop ]
%0 = phi i64 [ 2, %entry ], [ %fork, %loop ]
%inc1 = add i64 %0, 1
%inc2 = add i64 %0, 2
%inc11 = add i64 1, %inc1
%exact1 = ashr exact i64 %inc11, 32
%inc3 = add i64 %0, 3
%dummy_add = add i16 0, 0
%inc12 = add i64 1, %inc2
%exact2 = ashr exact i64 %inc12, 32
%dummy_shl = shl i64 %inc3, 32
%inc13 = add i64 1, %inc3
%exact3 = ashr exact i64 %inc13, 32
%fork = add i64 %0, 0
%sum1 = add i64 %exact3, %exact2
%sum2 = add i64 %sum1, %exact1
%zsum = add i64 %sum2, 0
%sext22 = add i64 1, %fork
%exact4 = ashr exact i64 %sext22, 32
%join = add i64 %fork, %zsum
%last = add i64 %join, %exact4
br label %loop
}

48
external/llvm/test/Transforms/SLPVectorizer/X86/PR35777.ll vendored
View File

@ -1,48 +0,0 @@
; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
; RUN: opt < %s -verify -slp-vectorizer -o - -S -mtriple=x86_64-apple-macosx10.13.0 | FileCheck %s
@global = local_unnamed_addr global [6 x double] zeroinitializer, align 16
define { i64, i64 } @patatino(double %arg) {
; CHECK-LABEL: @patatino(
; CHECK-NEXT: bb:
; CHECK-NEXT: [[TMP0:%.*]] = load <2 x double>, <2 x double>* bitcast ([6 x double]* @global to <2 x double>*), align 16
; CHECK-NEXT: [[TMP1:%.*]] = load <2 x double>, <2 x double>* bitcast (double* getelementptr inbounds ([6 x double], [6 x double]* @global, i64 0, i64 2) to <2 x double>*), align 16
; CHECK-NEXT: [[TMP2:%.*]] = insertelement <2 x double> undef, double [[ARG:%.*]], i32 0
; CHECK-NEXT: [[TMP3:%.*]] = insertelement <2 x double> [[TMP2]], double [[ARG]], i32 1
; CHECK-NEXT: [[TMP4:%.*]] = fmul <2 x double> [[TMP3]], [[TMP1]]
; CHECK-NEXT: [[TMP5:%.*]] = fadd <2 x double> [[TMP0]], [[TMP4]]
; CHECK-NEXT: [[TMP6:%.*]] = load <2 x double>, <2 x double>* bitcast (double* getelementptr inbounds ([6 x double], [6 x double]* @global, i64 0, i64 4) to <2 x double>*), align 16
; CHECK-NEXT: [[TMP7:%.*]] = fadd <2 x double> [[TMP6]], [[TMP5]]
; CHECK-NEXT: [[TMP8:%.*]] = fptosi <2 x double> [[TMP7]] to <2 x i32>
; CHECK-NEXT: [[TMP9:%.*]] = sext <2 x i32> [[TMP8]] to <2 x i64>
; CHECK-NEXT: [[TMP10:%.*]] = trunc <2 x i64> [[TMP9]] to <2 x i32>
; CHECK-NEXT: [[TMP11:%.*]] = extractelement <2 x i32> [[TMP10]], i32 0
; CHECK-NEXT: [[TMP12:%.*]] = sext i32 [[TMP11]] to i64
; CHECK-NEXT: [[TMP16:%.*]] = insertvalue { i64, i64 } undef, i64 [[TMP12]], 0
; CHECK-NEXT: [[TMP13:%.*]] = extractelement <2 x i32> [[TMP10]], i32 1
; CHECK-NEXT: [[TMP14:%.*]] = sext i32 [[TMP13]] to i64
; CHECK-NEXT: [[TMP17:%.*]] = insertvalue { i64, i64 } [[TMP16]], i64 [[TMP14]], 1
; CHECK-NEXT: ret { i64, i64 } [[TMP17]]
;
bb:
%tmp = load double, double* getelementptr inbounds ([6 x double], [6 x double]* @global, i64 0, i64 0), align 16
%tmp1 = load double, double* getelementptr inbounds ([6 x double], [6 x double]* @global, i64 0, i64 2), align 16
%tmp2 = fmul double %tmp1, %arg
%tmp3 = fadd double %tmp, %tmp2
%tmp4 = load double, double* getelementptr inbounds ([6 x double], [6 x double]* @global, i64 0, i64 4), align 16
%tmp5 = fadd double %tmp4, %tmp3
%tmp6 = fptosi double %tmp5 to i32
%tmp7 = sext i32 %tmp6 to i64
%tmp8 = load double, double* getelementptr inbounds ([6 x double], [6 x double]* @global, i64 0, i64 1), align 8
%tmp9 = load double, double* getelementptr inbounds ([6 x double], [6 x double]* @global, i64 0, i64 3), align 8
%tmp10 = fmul double %tmp9, %arg
%tmp11 = fadd double %tmp8, %tmp10
%tmp12 = load double, double* getelementptr inbounds ([6 x double], [6 x double]* @global, i64 0, i64 5), align 8
%tmp13 = fadd double %tmp12, %tmp11
%tmp14 = fptosi double %tmp13 to i32
%tmp15 = sext i32 %tmp14 to i64
%tmp16 = insertvalue { i64, i64 } undef, i64 %tmp7, 0
%tmp17 = insertvalue { i64, i64 } %tmp16, i64 %tmp15, 1
ret { i64, i64 } %tmp17
}

27
external/llvm/test/Transforms/SLPVectorizer/X86/PR35865.ll vendored
View File

@ -1,27 +0,0 @@
; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
; RUN: opt -slp-vectorizer < %s -S -o - -mtriple=x86_64-apple-macosx10.10.0 -mcpu=core2 | FileCheck %s
define void @_Z10fooConvertPDv4_xS0_S0_PKS_() {
; CHECK-LABEL: @_Z10fooConvertPDv4_xS0_S0_PKS_(
; CHECK-NEXT: entry:
; CHECK-NEXT: [[TMP0:%.*]] = extractelement <16 x half> undef, i32 4
; CHECK-NEXT: [[CONV_I_4_I:%.*]] = fpext half [[TMP0]] to float
; CHECK-NEXT: [[TMP1:%.*]] = bitcast float [[CONV_I_4_I]] to i32
; CHECK-NEXT: [[VECINS_I_4_I:%.*]] = insertelement <8 x i32> undef, i32 [[TMP1]], i32 4
; CHECK-NEXT: [[TMP2:%.*]] = extractelement <16 x half> undef, i32 5
; CHECK-NEXT: [[CONV_I_5_I:%.*]] = fpext half [[TMP2]] to float
; CHECK-NEXT: [[TMP3:%.*]] = bitcast float [[CONV_I_5_I]] to i32
; CHECK-NEXT: [[VECINS_I_5_I:%.*]] = insertelement <8 x i32> [[VECINS_I_4_I]], i32 [[TMP3]], i32 5
; CHECK-NEXT: ret void
;
entry:
%0 = extractelement <16 x half> undef, i32 4
%conv.i.4.i = fpext half %0 to float
%1 = bitcast float %conv.i.4.i to i32
%vecins.i.4.i = insertelement <8 x i32> undef, i32 %1, i32 4
%2 = extractelement <16 x half> undef, i32 5
%conv.i.5.i = fpext half %2 to float
%3 = bitcast float %conv.i.5.i to i32
%vecins.i.5.i = insertelement <8 x i32> %vecins.i.4.i, i32 %3, i32 5
ret void
}

314
external/llvm/test/Transforms/SLPVectorizer/X86/addsub.ll vendored
View File

@ -1,314 +0,0 @@
; RUN: opt < %s -basicaa -slp-vectorizer -S | FileCheck %s
target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
target triple = "x86_64-unknown-linux-gnu"
@b = common global [4 x i32] zeroinitializer, align 16
@c = common global [4 x i32] zeroinitializer, align 16
@d = common global [4 x i32] zeroinitializer, align 16
@e = common global [4 x i32] zeroinitializer, align 16
@a = common global [4 x i32] zeroinitializer, align 16
@fb = common global [4 x float] zeroinitializer, align 16
@fc = common global [4 x float] zeroinitializer, align 16
@fa = common global [4 x float] zeroinitializer, align 16
@fd = common global [4 x float] zeroinitializer, align 16
; CHECK-LABEL: @addsub
; CHECK: %5 = add nsw <4 x i32> %3, %4
; CHECK: %6 = add nsw <4 x i32> %2, %5
; CHECK: %7 = sub nsw <4 x i32> %2, %5
; CHECK: %8 = shufflevector <4 x i32> %6, <4 x i32> %7, <4 x i32> <i32 0, i32 5, i32 2, i32 7>
; Function Attrs: nounwind uwtable
define void @addsub() #0 {
entry:
%0 = load i32, i32* getelementptr inbounds ([4 x i32], [4 x i32]* @b, i32 0, i64 0), align 4
%1 = load i32, i32* getelementptr inbounds ([4 x i32], [4 x i32]* @c, i32 0, i64 0), align 4
%add = add nsw i32 %0, %1
%2 = load i32, i32* getelementptr inbounds ([4 x i32], [4 x i32]* @d, i32 0, i64 0), align 4
%3 = load i32, i32* getelementptr inbounds ([4 x i32], [4 x i32]* @e, i32 0, i64 0), align 4
%add1 = add nsw i32 %2, %3
%add2 = add nsw i32 %add, %add1
store i32 %add2, i32* getelementptr inbounds ([4 x i32], [4 x i32]* @a, i32 0, i64 0), align 4
%4 = load i32, i32* getelementptr inbounds ([4 x i32], [4 x i32]* @b, i32 0, i64 1), align 4
%5 = load i32, i32* getelementptr inbounds ([4 x i32], [4 x i32]* @c, i32 0, i64 1), align 4
%add3 = add nsw i32 %4, %5
%6 = load i32, i32* getelementptr inbounds ([4 x i32], [4 x i32]* @d, i32 0, i64 1), align 4
%7 = load i32, i32* getelementptr inbounds ([4 x i32], [4 x i32]* @e, i32 0, i64 1), align 4
%add4 = add nsw i32 %6, %7
%sub = sub nsw i32 %add3, %add4
store i32 %sub, i32* getelementptr inbounds ([4 x i32], [4 x i32]* @a, i32 0, i64 1), align 4
%8 = load i32, i32* getelementptr inbounds ([4 x i32], [4 x i32]* @b, i32 0, i64 2), align 4
%9 = load i32, i32* getelementptr inbounds ([4 x i32], [4 x i32]* @c, i32 0, i64 2), align 4
%add5 = add nsw i32 %8, %9
%10 = load i32, i32* getelementptr inbounds ([4 x i32], [4 x i32]* @d, i32 0, i64 2), align 4
%11 = load i32, i32* getelementptr inbounds ([4 x i32], [4 x i32]* @e, i32 0, i64 2), align 4
%add6 = add nsw i32 %10, %11
%add7 = add nsw i32 %add5, %add6
store i32 %add7, i32* getelementptr inbounds ([4 x i32], [4 x i32]* @a, i32 0, i64 2), align 4
%12 = load i32, i32* getelementptr inbounds ([4 x i32], [4 x i32]* @b, i32 0, i64 3), align 4
%13 = load i32, i32* getelementptr inbounds ([4 x i32], [4 x i32]* @c, i32 0, i64 3), align 4
%add8 = add nsw i32 %12, %13
%14 = load i32, i32* getelementptr inbounds ([4 x i32], [4 x i32]* @d, i32 0, i64 3), align 4
%15 = load i32, i32* getelementptr inbounds ([4 x i32], [4 x i32]* @e, i32 0, i64 3), align 4
%add9 = add nsw i32 %14, %15
%sub10 = sub nsw i32 %add8, %add9
store i32 %sub10, i32* getelementptr inbounds ([4 x i32], [4 x i32]* @a, i32 0, i64 3), align 4
ret void
}
; CHECK-LABEL: @subadd
; CHECK: %5 = add nsw <4 x i32> %3, %4
; CHECK: %6 = sub nsw <4 x i32> %2, %5
; CHECK: %7 = add nsw <4 x i32> %2, %5
; CHECK: %8 = shufflevector <4 x i32> %6, <4 x i32> %7, <4 x i32> <i32 0, i32 5, i32 2, i32 7>
; Function Attrs: nounwind uwtable
define void @subadd() #0 {
entry:
%0 = load i32, i32* getelementptr inbounds ([4 x i32], [4 x i32]* @b, i32 0, i64 0), align 4
%1 = load i32, i32* getelementptr inbounds ([4 x i32], [4 x i32]* @c, i32 0, i64 0), align 4
%add = add nsw i32 %0, %1
%2 = load i32, i32* getelementptr inbounds ([4 x i32], [4 x i32]* @d, i32 0, i64 0), align 4
%3 = load i32, i32* getelementptr inbounds ([4 x i32], [4 x i32]* @e, i32 0, i64 0), align 4
%add1 = add nsw i32 %2, %3
%sub = sub nsw i32 %add, %add1
store i32 %sub, i32* getelementptr inbounds ([4 x i32], [4 x i32]* @a, i32 0, i64 0), align 4
%4 = load i32, i32* getelementptr inbounds ([4 x i32], [4 x i32]* @b, i32 0, i64 1), align 4
%5 = load i32, i32* getelementptr inbounds ([4 x i32], [4 x i32]* @c, i32 0, i64 1), align 4
%add2 = add nsw i32 %4, %5
%6 = load i32, i32* getelementptr inbounds ([4 x i32], [4 x i32]* @d, i32 0, i64 1), align 4
%7 = load i32, i32* getelementptr inbounds ([4 x i32], [4 x i32]* @e, i32 0, i64 1), align 4
%add3 = add nsw i32 %6, %7
%add4 = add nsw i32 %add2, %add3
store i32 %add4, i32* getelementptr inbounds ([4 x i32], [4 x i32]* @a, i32 0, i64 1), align 4
%8 = load i32, i32* getelementptr inbounds ([4 x i32], [4 x i32]* @b, i32 0, i64 2), align 4
%9 = load i32, i32* getelementptr inbounds ([4 x i32], [4 x i32]* @c, i32 0, i64 2), align 4
%add5 = add nsw i32 %8, %9
%10 = load i32, i32* getelementptr inbounds ([4 x i32], [4 x i32]* @d, i32 0, i64 2), align 4
%11 = load i32, i32* getelementptr inbounds ([4 x i32], [4 x i32]* @e, i32 0, i64 2), align 4
%add6 = add nsw i32 %10, %11
%sub7 = sub nsw i32 %add5, %add6
store i32 %sub7, i32* getelementptr inbounds ([4 x i32], [4 x i32]* @a, i32 0, i64 2), align 4
%12 = load i32, i32* getelementptr inbounds ([4 x i32], [4 x i32]* @b, i32 0, i64 3), align 4
%13 = load i32, i32* getelementptr inbounds ([4 x i32], [4 x i32]* @c, i32 0, i64 3), align 4
%add8 = add nsw i32 %12, %13
%14 = load i32, i32* getelementptr inbounds ([4 x i32], [4 x i32]* @d, i32 0, i64 3), align 4
%15 = load i32, i32* getelementptr inbounds ([4 x i32], [4 x i32]* @e, i32 0, i64 3), align 4
%add9 = add nsw i32 %14, %15
%add10 = add nsw i32 %add8, %add9
store i32 %add10, i32* getelementptr inbounds ([4 x i32], [4 x i32]* @a, i32 0, i64 3), align 4
ret void
}
; CHECK-LABEL: @faddfsub
; CHECK: %2 = fadd <4 x float> %0, %1
; CHECK: %3 = fsub <4 x float> %0, %1
; CHECK: %4 = shufflevector <4 x float> %2, <4 x float> %3, <4 x i32> <i32 0, i32 5, i32 2, i32 7>
; Function Attrs: nounwind uwtable
define void @faddfsub() #0 {
entry:
%0 = load float, float* getelementptr inbounds ([4 x float], [4 x float]* @fb, i32 0, i64 0), align 4
%1 = load float, float* getelementptr inbounds ([4 x float], [4 x float]* @fc, i32 0, i64 0), align 4
%add = fadd float %0, %1
store float %add, float* getelementptr inbounds ([4 x float], [4 x float]* @fa, i32 0, i64 0), align 4
%2 = load float, float* getelementptr inbounds ([4 x float], [4 x float]* @fb, i32 0, i64 1), align 4
%3 = load float, float* getelementptr inbounds ([4 x float], [4 x float]* @fc, i32 0, i64 1), align 4
%sub = fsub float %2, %3
store float %sub, float* getelementptr inbounds ([4 x float], [4 x float]* @fa, i32 0, i64 1), align 4
%4 = load float, float* getelementptr inbounds ([4 x float], [4 x float]* @fb, i32 0, i64 2), align 4
%5 = load float, float* getelementptr inbounds ([4 x float], [4 x float]* @fc, i32 0, i64 2), align 4
%add1 = fadd float %4, %5
store float %add1, float* getelementptr inbounds ([4 x float], [4 x float]* @fa, i32 0, i64 2), align 4
%6 = load float, float* getelementptr inbounds ([4 x float], [4 x float]* @fb, i32 0, i64 3), align 4
%7 = load float, float* getelementptr inbounds ([4 x float], [4 x float]* @fc, i32 0, i64 3), align 4
%sub2 = fsub float %6, %7
store float %sub2, float* getelementptr inbounds ([4 x float], [4 x float]* @fa, i32 0, i64 3), align 4
ret void
}
; CHECK-LABEL: @fsubfadd
; CHECK: %2 = fsub <4 x float> %0, %1
; CHECK: %3 = fadd <4 x float> %0, %1
; CHECK: %4 = shufflevector <4 x float> %2, <4 x float> %3, <4 x i32> <i32 0, i32 5, i32 2, i32 7>
; Function Attrs: nounwind uwtable
define void @fsubfadd() #0 {
entry:
%0 = load float, float* getelementptr inbounds ([4 x float], [4 x float]* @fb, i32 0, i64 0), align 4
%1 = load float, float* getelementptr inbounds ([4 x float], [4 x float]* @fc, i32 0, i64 0), align 4
%sub = fsub float %0, %1
store float %sub, float* getelementptr inbounds ([4 x float], [4 x float]* @fa, i32 0, i64 0), align 4
%2 = load float, float* getelementptr inbounds ([4 x float], [4 x float]* @fb, i32 0, i64 1), align 4
%3 = load float, float* getelementptr inbounds ([4 x float], [4 x float]* @fc, i32 0, i64 1), align 4
%add = fadd float %2, %3
store float %add, float* getelementptr inbounds ([4 x float], [4 x float]* @fa, i32 0, i64 1), align 4
%4 = load float, float* getelementptr inbounds ([4 x float], [4 x float]* @fb, i32 0, i64 2), align 4
%5 = load float, float* getelementptr inbounds ([4 x float], [4 x float]* @fc, i32 0, i64 2), align 4
%sub1 = fsub float %4, %5
store float %sub1, float* getelementptr inbounds ([4 x float], [4 x float]* @fa, i32 0, i64 2), align 4
%6 = load float, float* getelementptr inbounds ([4 x float], [4 x float]* @fb, i32 0, i64 3), align 4
%7 = load float, float* getelementptr inbounds ([4 x float], [4 x float]* @fc, i32 0, i64 3), align 4
%add2 = fadd float %6, %7
store float %add2, float* getelementptr inbounds ([4 x float], [4 x float]* @fa, i32 0, i64 3), align 4
ret void
}
; CHECK-LABEL: @No_faddfsub
; CHECK-NOT: fadd <4 x float>
; CHECK-NOT: fsub <4 x float>
; CHECK-NOT: shufflevector
; Function Attrs: nounwind uwtable
define void @No_faddfsub() #0 {
entry:
%0 = load float, float* getelementptr inbounds ([4 x float], [4 x float]* @fb, i32 0, i64 0), align 4
%1 = load float, float* getelementptr inbounds ([4 x float], [4 x float]* @fc, i32 0, i64 0), align 4
%add = fadd float %0, %1
store float %add, float* getelementptr inbounds ([4 x float], [4 x float]* @fa, i32 0, i64 0), align 4
%2 = load float, float* getelementptr inbounds ([4 x float], [4 x float]* @fb, i32 0, i64 1), align 4
%3 = load float, float* getelementptr inbounds ([4 x float], [4 x float]* @fc, i32 0, i64 1), align 4
%add1 = fadd float %2, %3
store float %add1, float* getelementptr inbounds ([4 x float], [4 x float]* @fa, i32 0, i64 1), align 4
%4 = load float, float* getelementptr inbounds ([4 x float], [4 x float]* @fb, i32 0, i64 2), align 4
%5 = load float, float* getelementptr inbounds ([4 x float], [4 x float]* @fc, i32 0, i64 2), align 4
%add2 = fadd float %4, %5
store float %add2, float* getelementptr inbounds ([4 x float], [4 x float]* @fa, i32 0, i64 2), align 4
%6 = load float, float* getelementptr inbounds ([4 x float], [4 x float]* @fb, i32 0, i64 3), align 4
%7 = load float, float* getelementptr inbounds ([4 x float], [4 x float]* @fc, i32 0, i64 3), align 4
%sub = fsub float %6, %7
store float %sub, float* getelementptr inbounds ([4 x float], [4 x float]* @fa, i32 0, i64 3), align 4
ret void
}
; Check vectorization of following code for float data type-
; fc[0] = fb[0]+fa[0]; //swapped fb and fa
; fc[1] = fa[1]-fb[1];
; fc[2] = fa[2]+fb[2];
; fc[3] = fa[3]-fb[3];
; CHECK-LABEL: @reorder_alt
; CHECK: %3 = fadd <4 x float> %1, %2
; CHECK: %4 = fsub <4 x float> %1, %2
; CHECK: %5 = shufflevector <4 x float> %3, <4 x float> %4, <4 x i32> <i32 0, i32 5, i32 2, i32 7>
define void @reorder_alt() #0 {
%1 = load float, float* getelementptr inbounds ([4 x float], [4 x float]* @fb, i32 0, i64 0), align 4
%2 = load float, float* getelementptr inbounds ([4 x float], [4 x float]* @fa, i32 0, i64 0), align 4
%3 = fadd float %1, %2
store float %3, float* getelementptr inbounds ([4 x float], [4 x float]* @fc, i32 0, i64 0), align 4
%4 = load float, float* getelementptr inbounds ([4 x float], [4 x float]* @fa, i32 0, i64 1), align 4
%5 = load float, float* getelementptr inbounds ([4 x float], [4 x float]* @fb, i32 0, i64 1), align 4
%6 = fsub float %4, %5
store float %6, float* getelementptr inbounds ([4 x float], [4 x float]* @fc, i32 0, i64 1), align 4
%7 = load float, float* getelementptr inbounds ([4 x float], [4 x float]* @fa, i32 0, i64 2), align 4
%8 = load float, float* getelementptr inbounds ([4 x float], [4 x float]* @fb, i32 0, i64 2), align 4
%9 = fadd float %7, %8
store float %9, float* getelementptr inbounds ([4 x float], [4 x float]* @fc, i32 0, i64 2), align 4
%10 = load float, float* getelementptr inbounds ([4 x float], [4 x float]* @fa, i32 0, i64 3), align 4
%11 = load float, float* getelementptr inbounds ([4 x float], [4 x float]* @fb, i32 0, i64 3), align 4
%12 = fsub float %10, %11
store float %12, float* getelementptr inbounds ([4 x float], [4 x float]* @fc, i32 0, i64 3), align 4
ret void
}
; Check vectorization of following code for float data type-
; fc[0] = fa[0]+(fb[0]-fd[0]);
; fc[1] = fa[1]-(fb[1]+fd[1]);
; fc[2] = fa[2]+(fb[2]-fd[2]);
; fc[3] = fa[3]-(fd[3]+fb[3]); //swapped fd and fb
; CHECK-LABEL: @reorder_alt_subTree
; CHECK: %4 = fsub <4 x float> %3, %2
; CHECK: %5 = fadd <4 x float> %3, %2
; CHECK: %6 = shufflevector <4 x float> %4, <4 x float> %5, <4 x i32> <i32 0, i32 5, i32 2, i32 7>
; CHECK: %7 = fadd <4 x float> %1, %6
; CHECK: %8 = fsub <4 x float> %1, %6
; CHECK: %9 = shufflevector <4 x float> %7, <4 x float> %8, <4 x i32> <i32 0, i32 5, i32 2, i32 7>
define void @reorder_alt_subTree() #0 {
%1 = load float, float* getelementptr inbounds ([4 x float], [4 x float]* @fa, i32 0, i64 0), align 4
%2 = load float, float* getelementptr inbounds ([4 x float], [4 x float]* @fb, i32 0, i64 0), align 4
%3 = load float, float* getelementptr inbounds ([4 x float], [4 x float]* @fd, i32 0, i64 0), align 4
%4 = fsub float %2, %3
%5 = fadd float %1, %4
store float %5, float* getelementptr inbounds ([4 x float], [4 x float]* @fc, i32 0, i64 0), align 4
%6 = load float, float* getelementptr inbounds ([4 x float], [4 x float]* @fa, i32 0, i64 1), align 4
%7 = load float, float* getelementptr inbounds ([4 x float], [4 x float]* @fb, i32 0, i64 1), align 4
%8 = load float, float* getelementptr inbounds ([4 x float], [4 x float]* @fd, i32 0, i64 1), align 4
%9 = fadd float %7, %8
%10 = fsub float %6, %9
store float %10, float* getelementptr inbounds ([4 x float], [4 x float]* @fc, i32 0, i64 1), align 4
%11 = load float, float* getelementptr inbounds ([4 x float], [4 x float]* @fa, i32 0, i64 2), align 4
%12 = load float, float* getelementptr inbounds ([4 x float], [4 x float]* @fb, i32 0, i64 2), align 4
%13 = load float, float* getelementptr inbounds ([4 x float], [4 x float]* @fd, i32 0, i64 2), align 4
%14 = fsub float %12, %13
%15 = fadd float %11, %14
store float %15, float* getelementptr inbounds ([4 x float], [4 x float]* @fc, i32 0, i64 2), align 4
%16 = load float, float* getelementptr inbounds ([4 x float], [4 x float]* @fa, i32 0, i64 3), align 4
%17 = load float, float* getelementptr inbounds ([4 x float], [4 x float]* @fd, i32 0, i64 3), align 4
%18 = load float, float* getelementptr inbounds ([4 x float], [4 x float]* @fb, i32 0, i64 3), align 4
%19 = fadd float %17, %18
%20 = fsub float %16, %19
store float %20, float* getelementptr inbounds ([4 x float], [4 x float]* @fc, i32 0, i64 3), align 4
ret void
}
; Check vectorization of following code for double data type-
; c[0] = (a[0]+b[0])-d[0];
; c[1] = d[1]+(a[1]+b[1]); //swapped d[1] and (a[1]+b[1])
; CHECK-LABEL: @reorder_alt_rightsubTree
; CHECK: fadd <2 x double>
; CHECK: fsub <2 x double>
; CHECK: shufflevector <2 x double>
define void @reorder_alt_rightsubTree(double* nocapture %c, double* noalias nocapture readonly %a, double* noalias nocapture readonly %b, double* noalias nocapture readonly %d) {
%1 = load double, double* %a
%2 = load double, double* %b
%3 = fadd double %1, %2
%4 = load double, double* %d
%5 = fsub double %3, %4
store double %5, double* %c
%6 = getelementptr inbounds double, double* %d, i64 1
%7 = load double, double* %6
%8 = getelementptr inbounds double, double* %a, i64 1
%9 = load double, double* %8
%10 = getelementptr inbounds double, double* %b, i64 1
%11 = load double, double* %10
%12 = fadd double %9, %11
%13 = fadd double %7, %12
%14 = getelementptr inbounds double, double* %c, i64 1
store double %13, double* %14
ret void
}
; Dont vectorization of following code for float data type as sub is not commutative-
; fc[0] = fb[0]+fa[0];
; fc[1] = fa[1]-fb[1];
; fc[2] = fa[2]+fb[2];
; fc[3] = fb[3]-fa[3];
; In the above code we can swap the 1st and 2nd operation as fadd is commutative
; but not 2nd or 4th as fsub is not commutative.
; CHECK-LABEL: @no_vec_shuff_reorder
; CHECK-NOT: fadd <4 x float>
; CHECK-NOT: fsub <4 x float>
; CHECK-NOT: shufflevector
define void @no_vec_shuff_reorder() #0 {
%1 = load float, float* getelementptr inbounds ([4 x float], [4 x float]* @fb, i32 0, i64 0), align 4
%2 = load float, float* getelementptr inbounds ([4 x float], [4 x float]* @fa, i32 0, i64 0), align 4
%3 = fadd float %1, %2
store float %3, float* getelementptr inbounds ([4 x float], [4 x float]* @fc, i32 0, i64 0), align 4
%4 = load float, float* getelementptr inbounds ([4 x float], [4 x float]* @fa, i32 0, i64 1), align 4
%5 = load float, float* getelementptr inbounds ([4 x float], [4 x float]* @fb, i32 0, i64 1), align 4
%6 = fsub float %4, %5
store float %6, float* getelementptr inbounds ([4 x float], [4 x float]* @fc, i32 0, i64 1), align 4
%7 = load float, float* getelementptr inbounds ([4 x float], [4 x float]* @fa, i32 0, i64 2), align 4
%8 = load float, float* getelementptr inbounds ([4 x float], [4 x float]* @fb, i32 0, i64 2), align 4
%9 = fadd float %7, %8
store float %9, float* getelementptr inbounds ([4 x float], [4 x float]* @fc, i32 0, i64 2), align 4
%10 = load float, float* getelementptr inbounds ([4 x float], [4 x float]* @fb, i32 0, i64 3), align 4
%11 = load float, float* getelementptr inbounds ([4 x float], [4 x float]* @fa, i32 0, i64 3), align 4
%12 = fsub float %10, %11
store float %12, float* getelementptr inbounds ([4 x float], [4 x float]* @fc, i32 0, i64 3), align 4
ret void
}
attributes #0 = { nounwind }

19
external/llvm/test/Transforms/SLPVectorizer/X86/aggregate.ll vendored
View File

@ -1,19 +0,0 @@
; RUN: opt -S -mtriple=x86_64-unknown-linux -mcpu=corei7 -slp-vectorizer < %s | FileCheck %s
%struct.S = type { i8*, i8* }
@kS0 = common global %struct.S zeroinitializer, align 8
define { i64, i64 } @getS() {
entry:
%0 = load i64, i64* bitcast (%struct.S* @kS0 to i64*), align 8
%1 = load i64, i64* bitcast (i8** getelementptr inbounds (%struct.S, %struct.S* @kS0, i64 0, i32 1) to i64*), align 8
%2 = insertvalue { i64, i64 } undef, i64 %0, 0
%3 = insertvalue { i64, i64 } %2, i64 %1, 1
ret { i64, i64 } %3
}
; CHECK: load i64
; CHECK-NOT: load <2 x i64>
; CHECK-NOT: extractelement

55
external/llvm/test/Transforms/SLPVectorizer/X86/align.ll vendored
View File

@ -1,55 +0,0 @@
; RUN: opt < %s -basicaa -slp-vectorizer -S -mtriple=x86_64-apple-macosx10.8.0 -mcpu=corei7-avx | FileCheck %s
target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64-S128"
target triple = "x86_64-apple-macosx10.8.0"
; Simple 3-pair chain with loads and stores
; CHECK-LABEL: @test1
define void @test1(double* %a, double* %b, double* %c) {
entry:
%agg.tmp.i.i.sroa.0 = alloca [3 x double], align 16
; CHECK: %[[V0:[0-9]+]] = load <2 x double>, <2 x double>* %[[V2:[0-9]+]], align 8
%i0 = load double, double* %a
%i1 = load double, double* %b
%mul = fmul double %i0, %i1
%store1 = getelementptr inbounds [3 x double], [3 x double]* %agg.tmp.i.i.sroa.0, i64 0, i64 1
%store2 = getelementptr inbounds [3 x double], [3 x double]* %agg.tmp.i.i.sroa.0, i64 0, i64 2
%arrayidx3 = getelementptr inbounds double, double* %a, i64 1
%i3 = load double, double* %arrayidx3, align 8
%arrayidx4 = getelementptr inbounds double, double* %b, i64 1
%i4 = load double, double* %arrayidx4, align 8
%mul5 = fmul double %i3, %i4
; CHECK: store <2 x double> %[[V1:[0-9]+]], <2 x double>* %[[V2:[0-9]+]], align 8
store double %mul, double* %store1
store double %mul5, double* %store2, align 16
; CHECK: ret
ret void
}
; Float has 4 byte abi alignment on x86_64. We must use the alignmnet of the
; value being loaded/stored not the alignment of the pointer type.
; CHECK-LABEL: @test2
; CHECK-NOT: align 8
; CHECK: load <4 x float>{{.*}}, align 4
; CHECK: store <4 x float>{{.*}}, align 4
; CHECK: ret
define void @test2(float * %a, float * %b) {
entry:
%l0 = load float, float* %a
%a1 = getelementptr inbounds float, float* %a, i64 1
%l1 = load float, float* %a1
%a2 = getelementptr inbounds float, float* %a, i64 2
%l2 = load float, float* %a2
%a3 = getelementptr inbounds float, float* %a, i64 3
%l3 = load float, float* %a3
store float %l0, float* %b
%b1 = getelementptr inbounds float, float* %b, i64 1
store float %l1, float* %b1
%b2 = getelementptr inbounds float, float* %b, i64 2
store float %l2, float* %b2
%b3 = getelementptr inbounds float, float* %b, i64 3
store float %l3, float* %b3
ret void
}

707
external/llvm/test/Transforms/SLPVectorizer/X86/arith-add.ll vendored
View File

File diff suppressed because it is too large Load Diff

1323
external/llvm/test/Transforms/SLPVectorizer/X86/arith-fp.ll vendored
View File

File diff suppressed because it is too large Load Diff

774
external/llvm/test/Transforms/SLPVectorizer/X86/arith-mul.ll vendored
View File

File diff suppressed because it is too large Load Diff

707
external/llvm/test/Transforms/SLPVectorizer/X86/arith-sub.ll vendored
View File

File diff suppressed because it is too large Load Diff

31
external/llvm/test/Transforms/SLPVectorizer/X86/atomics.ll vendored
View File

@ -1,31 +0,0 @@
; RUN: opt < %s -basicaa -slp-vectorizer -S |FileCheck %s
target datalayout = "e-m:o-i64:64-f80:128-n8:16:32:64-S128"
@x = global [4 x i32] zeroinitializer, align 16
@a = global [4 x i32] zeroinitializer, align 16
; The SLPVectorizer should not vectorize atomic stores and it should not
; schedule regular stores around atomic stores.
; CHECK-LABEL: test
; CHECK: store i32
; CHECK: store atomic i32
; CHECK: store i32
; CHECK: store atomic i32
; CHECK: store i32
; CHECK: store atomic i32
; CHECK: store i32
; CHECK: store atomic i32
define void @test() {
entry:
store i32 0, i32* getelementptr inbounds ([4 x i32], [4 x i32]* @a, i64 0, i64 0), align 16
store atomic i32 0, i32* getelementptr inbounds ([4 x i32], [4 x i32]* @x, i64 0, i64 0) release, align 16
store i32 0, i32* getelementptr inbounds ([4 x i32], [4 x i32]* @a, i64 0, i64 1), align 4
store atomic i32 1, i32* getelementptr inbounds ([4 x i32], [4 x i32]* @x, i64 0, i64 1) release, align 4
store i32 0, i32* getelementptr inbounds ([4 x i32], [4 x i32]* @a, i64 0, i64 2), align 8
store atomic i32 2, i32* getelementptr inbounds ([4 x i32], [4 x i32]* @x, i64 0, i64 2) release, align 8
store i32 0, i32* getelementptr inbounds ([4 x i32], [4 x i32]* @a, i64 0, i64 3), align 4
store atomic i32 3, i32* getelementptr inbounds ([4 x i32], [4 x i32]* @x, i64 0, i64 3) release, align 4
ret void
}

76
external/llvm/test/Transforms/SLPVectorizer/X86/bad_types.ll vendored
View File

@ -1,76 +0,0 @@
; RUN: opt < %s -basicaa -slp-vectorizer -S -mcpu=corei7-avx | FileCheck %s
target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
target triple = "x86_64-unknown-linux-gnu"
define void @test1(x86_mmx %a, x86_mmx %b, i64* %ptr) {
; Ensure we can handle x86_mmx values which are primitive and can be bitcast
; with integer types but can't be put into a vector.
;
; CHECK-LABEL: @test1
; CHECK: store i64
; CHECK: store i64
; CHECK: ret void
entry:
%a.cast = bitcast x86_mmx %a to i64
%b.cast = bitcast x86_mmx %b to i64
%a.and = and i64 %a.cast, 42
%b.and = and i64 %b.cast, 42
%gep = getelementptr i64, i64* %ptr, i32 1
store i64 %a.and, i64* %ptr
store i64 %b.and, i64* %gep
ret void
}
define void @test2(x86_mmx %a, x86_mmx %b) {
; Same as @test1 but using phi-input vectorization instead of store
; vectorization.
;
; CHECK-LABEL: @test2
; CHECK: and i64
; CHECK: and i64
; CHECK: ret void
entry:
br i1 undef, label %if.then, label %exit
if.then:
%a.cast = bitcast x86_mmx %a to i64
%b.cast = bitcast x86_mmx %b to i64
%a.and = and i64 %a.cast, 42
%b.and = and i64 %b.cast, 42
br label %exit
exit:
%a.phi = phi i64 [ 0, %entry ], [ %a.and, %if.then ]
%b.phi = phi i64 [ 0, %entry ], [ %b.and, %if.then ]
tail call void @f(i64 %a.phi, i64 %b.phi)
ret void
}
define i8 @test3(i8 *%addr) {
; Check that we do not vectorize types that are padded to a bigger ones.
;
; CHECK-LABEL: @test3
; CHECK-NOT: <4 x i2>
; CHECK: ret i8
entry:
%a = bitcast i8* %addr to i2*
%a0 = getelementptr inbounds i2, i2* %a, i64 0
%a1 = getelementptr inbounds i2, i2* %a, i64 1
%a2 = getelementptr inbounds i2, i2* %a, i64 2
%a3 = getelementptr inbounds i2, i2* %a, i64 3
%l0 = load i2, i2* %a0, align 1
%l1 = load i2, i2* %a1, align 1
%l2 = load i2, i2* %a2, align 1
%l3 = load i2, i2* %a3, align 1
br label %bb1
bb1: ; preds = %entry
%p0 = phi i2 [ %l0, %entry ]
%p1 = phi i2 [ %l1, %entry ]
%p2 = phi i2 [ %l2, %entry ]
%p3 = phi i2 [ %l3, %entry ]
%r = zext i2 %p2 to i8
ret i8 %r
}
declare void @f(i64, i64)

32
external/llvm/test/Transforms/SLPVectorizer/X86/barriercall.ll vendored
View File

@ -1,32 +0,0 @@
; RUN: opt < %s -basicaa -slp-vectorizer -dce -S -mtriple=x86_64-apple-macosx10.8.0 -mcpu=corei7-avx | FileCheck %s
target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64-S128"
target triple = "x86_64-apple-macosx10.8.0"
;CHECK-LABEL: @foo(
;CHECK: store <4 x i32>
;CHECK: ret
define i32 @foo(i32* nocapture %A, i32 %n) {
entry:
%call = tail call i32 (...) @bar() #2
%mul = mul nsw i32 %n, 5
%add = add nsw i32 %mul, 9
store i32 %add, i32* %A, align 4
%mul1 = mul nsw i32 %n, 9
%add2 = add nsw i32 %mul1, 9
%arrayidx3 = getelementptr inbounds i32, i32* %A, i64 1
store i32 %add2, i32* %arrayidx3, align 4
%mul4 = shl i32 %n, 3
%add5 = add nsw i32 %mul4, 9
%arrayidx6 = getelementptr inbounds i32, i32* %A, i64 2
store i32 %add5, i32* %arrayidx6, align 4
%mul7 = mul nsw i32 %n, 10
%add8 = add nsw i32 %mul7, 9
%arrayidx9 = getelementptr inbounds i32, i32* %A, i64 3
store i32 %add8, i32* %arrayidx9, align 4
ret i32 undef
}
; We can still vectorize the stores below.
declare i32 @bar(...)

423
external/llvm/test/Transforms/SLPVectorizer/X86/bitreverse.ll vendored
View File

@ -1,423 +0,0 @@
; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
; RUN: opt < %s -mtriple=x86_64-unknown -basicaa -slp-vectorizer -S | FileCheck %s --check-prefix=CHECK --check-prefix=SSE
; RUN: opt < %s -mtriple=x86_64-unknown -mcpu=corei7-avx -basicaa -slp-vectorizer -S | FileCheck %s --check-prefix=CHECK --check-prefix=AVX --check-prefix=AVX1
; RUN: opt < %s -mtriple=x86_64-unknown -mcpu=core-avx2 -basicaa -slp-vectorizer -S | FileCheck %s --check-prefix=CHECK --check-prefix=AVX --check-prefix=AVX2
; RUN: opt < %s -mtriple=x86_64-unknown -mcpu=bdver2 -basicaa -slp-vectorizer -S | FileCheck %s --check-prefix=CHECK --check-prefix=XOP
; RUN: opt < %s -mtriple=x86_64-unknown -mcpu=bdver4 -basicaa -slp-vectorizer -S | FileCheck %s --check-prefix=CHECK --check-prefix=XOP
target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
@src64 = common global [4 x i64] zeroinitializer, align 32
@dst64 = common global [4 x i64] zeroinitializer, align 32
@src32 = common global [8 x i32] zeroinitializer, align 32
@dst32 = common global [8 x i32] zeroinitializer, align 32
@src16 = common global [16 x i16] zeroinitializer, align 32
@dst16 = common global [16 x i16] zeroinitializer, align 32
@src8 = common global [32 x i8] zeroinitializer, align 32
@dst8 = common global [32 x i8] zeroinitializer, align 32
declare i64 @llvm.bitreverse.i64(i64)
declare i32 @llvm.bitreverse.i32(i32)
declare i16 @llvm.bitreverse.i16(i16)
declare i8 @llvm.bitreverse.i8(i8)
define void @bitreverse_2i64() #0 {
; CHECK-LABEL: @bitreverse_2i64(
; CHECK-NEXT: [[TMP1:%.*]] = load <2 x i64>, <2 x i64>* bitcast ([4 x i64]* @src64 to <2 x i64>*), align 8
; CHECK-NEXT: [[TMP2:%.*]] = call <2 x i64> @llvm.bitreverse.v2i64(<2 x i64> [[TMP1]])
; CHECK-NEXT: store <2 x i64> [[TMP2]], <2 x i64>* bitcast ([4 x i64]* @dst64 to <2 x i64>*), align 8
; CHECK-NEXT: ret void
;
%ld0 = load i64, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @src64, i32 0, i64 0), align 8
%ld1 = load i64, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @src64, i32 0, i64 1), align 8
%bitreverse0 = call i64 @llvm.bitreverse.i64(i64 %ld0)
%bitreverse1 = call i64 @llvm.bitreverse.i64(i64 %ld1)
store i64 %bitreverse0, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @dst64, i32 0, i64 0), align 8
store i64 %bitreverse1, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @dst64, i32 0, i64 1), align 8
ret void
}
define void @bitreverse_4i64() #0 {
; SSE-LABEL: @bitreverse_4i64(
; SSE-NEXT: [[TMP1:%.*]] = load <2 x i64>, <2 x i64>* bitcast ([4 x i64]* @src64 to <2 x i64>*), align 4
; SSE-NEXT: [[TMP2:%.*]] = load <2 x i64>, <2 x i64>* bitcast (i64* getelementptr inbounds ([4 x i64], [4 x i64]* @src64, i64 0, i64 2) to <2 x i64>*), align 4
; SSE-NEXT: [[TMP3:%.*]] = call <2 x i64> @llvm.bitreverse.v2i64(<2 x i64> [[TMP1]])
; SSE-NEXT: [[TMP4:%.*]] = call <2 x i64> @llvm.bitreverse.v2i64(<2 x i64> [[TMP2]])
; SSE-NEXT: store <2 x i64> [[TMP3]], <2 x i64>* bitcast ([4 x i64]* @dst64 to <2 x i64>*), align 4
; SSE-NEXT: store <2 x i64> [[TMP4]], <2 x i64>* bitcast (i64* getelementptr inbounds ([4 x i64], [4 x i64]* @dst64, i64 0, i64 2) to <2 x i64>*), align 4
; SSE-NEXT: ret void
;
; AVX-LABEL: @bitreverse_4i64(
; AVX-NEXT: [[TMP1:%.*]] = load <4 x i64>, <4 x i64>* bitcast ([4 x i64]* @src64 to <4 x i64>*), align 4
; AVX-NEXT: [[TMP2:%.*]] = call <4 x i64> @llvm.bitreverse.v4i64(<4 x i64> [[TMP1]])
; AVX-NEXT: store <4 x i64> [[TMP2]], <4 x i64>* bitcast ([4 x i64]* @dst64 to <4 x i64>*), align 4
; AVX-NEXT: ret void
;
; XOP-LABEL: @bitreverse_4i64(
; XOP-NEXT: [[TMP1:%.*]] = load <4 x i64>, <4 x i64>* bitcast ([4 x i64]* @src64 to <4 x i64>*), align 4
; XOP-NEXT: [[TMP2:%.*]] = call <4 x i64> @llvm.bitreverse.v4i64(<4 x i64> [[TMP1]])
; XOP-NEXT: store <4 x i64> [[TMP2]], <4 x i64>* bitcast ([4 x i64]* @dst64 to <4 x i64>*), align 4
; XOP-NEXT: ret void
;
%ld0 = load i64, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @src64, i64 0, i64 0), align 4
%ld1 = load i64, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @src64, i64 0, i64 1), align 4
%ld2 = load i64, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @src64, i64 0, i64 2), align 4
%ld3 = load i64, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @src64, i64 0, i64 3), align 4
%bitreverse0 = call i64 @llvm.bitreverse.i64(i64 %ld0)
%bitreverse1 = call i64 @llvm.bitreverse.i64(i64 %ld1)
%bitreverse2 = call i64 @llvm.bitreverse.i64(i64 %ld2)
%bitreverse3 = call i64 @llvm.bitreverse.i64(i64 %ld3)
store i64 %bitreverse0, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @dst64, i64 0, i64 0), align 4
store i64 %bitreverse1, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @dst64, i64 0, i64 1), align 4
store i64 %bitreverse2, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @dst64, i64 0, i64 2), align 4
store i64 %bitreverse3, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @dst64, i64 0, i64 3), align 4
ret void
}
define void @bitreverse_4i32() #0 {
; CHECK-LABEL: @bitreverse_4i32(
; CHECK-NEXT: [[TMP1:%.*]] = load <4 x i32>, <4 x i32>* bitcast ([8 x i32]* @src32 to <4 x i32>*), align 4
; CHECK-NEXT: [[TMP2:%.*]] = call <4 x i32> @llvm.bitreverse.v4i32(<4 x i32> [[TMP1]])
; CHECK-NEXT: store <4 x i32> [[TMP2]], <4 x i32>* bitcast ([8 x i32]* @dst32 to <4 x i32>*), align 4
; CHECK-NEXT: ret void
;
%ld0 = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 0), align 4
%ld1 = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 1), align 4
%ld2 = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 2), align 4
%ld3 = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 3), align 4
%bitreverse0 = call i32 @llvm.bitreverse.i32(i32 %ld0)
%bitreverse1 = call i32 @llvm.bitreverse.i32(i32 %ld1)
%bitreverse2 = call i32 @llvm.bitreverse.i32(i32 %ld2)
%bitreverse3 = call i32 @llvm.bitreverse.i32(i32 %ld3)
store i32 %bitreverse0, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 0), align 4
store i32 %bitreverse1, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 1), align 4
store i32 %bitreverse2, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 2), align 4
store i32 %bitreverse3, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 3), align 4
ret void
}
define void @bitreverse_8i32() #0 {
; SSE-LABEL: @bitreverse_8i32(
; SSE-NEXT: [[TMP1:%.*]] = load <4 x i32>, <4 x i32>* bitcast ([8 x i32]* @src32 to <4 x i32>*), align 2
; SSE-NEXT: [[TMP2:%.*]] = load <4 x i32>, <4 x i32>* bitcast (i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 4) to <4 x i32>*), align 2
; SSE-NEXT: [[TMP3:%.*]] = call <4 x i32> @llvm.bitreverse.v4i32(<4 x i32> [[TMP1]])
; SSE-NEXT: [[TMP4:%.*]] = call <4 x i32> @llvm.bitreverse.v4i32(<4 x i32> [[TMP2]])
; SSE-NEXT: store <4 x i32> [[TMP3]], <4 x i32>* bitcast ([8 x i32]* @dst32 to <4 x i32>*), align 2
; SSE-NEXT: store <4 x i32> [[TMP4]], <4 x i32>* bitcast (i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 4) to <4 x i32>*), align 2
; SSE-NEXT: ret void
;
; AVX-LABEL: @bitreverse_8i32(
; AVX-NEXT: [[TMP1:%.*]] = load <8 x i32>, <8 x i32>* bitcast ([8 x i32]* @src32 to <8 x i32>*), align 2
; AVX-NEXT: [[TMP2:%.*]] = call <8 x i32> @llvm.bitreverse.v8i32(<8 x i32> [[TMP1]])
; AVX-NEXT: store <8 x i32> [[TMP2]], <8 x i32>* bitcast ([8 x i32]* @dst32 to <8 x i32>*), align 2
; AVX-NEXT: ret void
;
; XOP-LABEL: @bitreverse_8i32(
; XOP-NEXT: [[TMP1:%.*]] = load <8 x i32>, <8 x i32>* bitcast ([8 x i32]* @src32 to <8 x i32>*), align 2
; XOP-NEXT: [[TMP2:%.*]] = call <8 x i32> @llvm.bitreverse.v8i32(<8 x i32> [[TMP1]])
; XOP-NEXT: store <8 x i32> [[TMP2]], <8 x i32>* bitcast ([8 x i32]* @dst32 to <8 x i32>*), align 2
; XOP-NEXT: ret void
;
%ld0 = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 0), align 2
%ld1 = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 1), align 2
%ld2 = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 2), align 2
%ld3 = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 3), align 2
%ld4 = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 4), align 2
%ld5 = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 5), align 2
%ld6 = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 6), align 2
%ld7 = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 7), align 2
%bitreverse0 = call i32 @llvm.bitreverse.i32(i32 %ld0)
%bitreverse1 = call i32 @llvm.bitreverse.i32(i32 %ld1)
%bitreverse2 = call i32 @llvm.bitreverse.i32(i32 %ld2)
%bitreverse3 = call i32 @llvm.bitreverse.i32(i32 %ld3)
%bitreverse4 = call i32 @llvm.bitreverse.i32(i32 %ld4)
%bitreverse5 = call i32 @llvm.bitreverse.i32(i32 %ld5)
%bitreverse6 = call i32 @llvm.bitreverse.i32(i32 %ld6)
%bitreverse7 = call i32 @llvm.bitreverse.i32(i32 %ld7)
store i32 %bitreverse0, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 0), align 2
store i32 %bitreverse1, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 1), align 2
store i32 %bitreverse2, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 2), align 2
store i32 %bitreverse3, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 3), align 2
store i32 %bitreverse4, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 4), align 2
store i32 %bitreverse5, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 5), align 2
store i32 %bitreverse6, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 6), align 2
store i32 %bitreverse7, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 7), align 2
ret void
}
define void @bitreverse_8i16() #0 {
; CHECK-LABEL: @bitreverse_8i16(
; CHECK-NEXT: [[TMP1:%.*]] = load <8 x i16>, <8 x i16>* bitcast ([16 x i16]* @src16 to <8 x i16>*), align 2
; CHECK-NEXT: [[TMP2:%.*]] = call <8 x i16> @llvm.bitreverse.v8i16(<8 x i16> [[TMP1]])
; CHECK-NEXT: store <8 x i16> [[TMP2]], <8 x i16>* bitcast ([16 x i16]* @dst16 to <8 x i16>*), align 2
; CHECK-NEXT: ret void
;
%ld0 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 0), align 2
%ld1 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 1), align 2
%ld2 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 2), align 2
%ld3 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 3), align 2
%ld4 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 4), align 2
%ld5 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 5), align 2
%ld6 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 6), align 2
%ld7 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 7), align 2
%bitreverse0 = call i16 @llvm.bitreverse.i16(i16 %ld0)
%bitreverse1 = call i16 @llvm.bitreverse.i16(i16 %ld1)
%bitreverse2 = call i16 @llvm.bitreverse.i16(i16 %ld2)
%bitreverse3 = call i16 @llvm.bitreverse.i16(i16 %ld3)
%bitreverse4 = call i16 @llvm.bitreverse.i16(i16 %ld4)
%bitreverse5 = call i16 @llvm.bitreverse.i16(i16 %ld5)
%bitreverse6 = call i16 @llvm.bitreverse.i16(i16 %ld6)
%bitreverse7 = call i16 @llvm.bitreverse.i16(i16 %ld7)
store i16 %bitreverse0, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 0), align 2
store i16 %bitreverse1, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 1), align 2
store i16 %bitreverse2, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 2), align 2
store i16 %bitreverse3, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 3), align 2
store i16 %bitreverse4, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 4), align 2
store i16 %bitreverse5, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 5), align 2
store i16 %bitreverse6, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 6), align 2
store i16 %bitreverse7, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 7), align 2
ret void
}
define void @bitreverse_16i16() #0 {
; SSE-LABEL: @bitreverse_16i16(
; SSE-NEXT: [[TMP1:%.*]] = load <8 x i16>, <8 x i16>* bitcast ([16 x i16]* @src16 to <8 x i16>*), align 2
; SSE-NEXT: [[TMP2:%.*]] = load <8 x i16>, <8 x i16>* bitcast (i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 8) to <8 x i16>*), align 2
; SSE-NEXT: [[TMP3:%.*]] = call <8 x i16> @llvm.bitreverse.v8i16(<8 x i16> [[TMP1]])
; SSE-NEXT: [[TMP4:%.*]] = call <8 x i16> @llvm.bitreverse.v8i16(<8 x i16> [[TMP2]])
; SSE-NEXT: store <8 x i16> [[TMP3]], <8 x i16>* bitcast ([16 x i16]* @dst16 to <8 x i16>*), align 2
; SSE-NEXT: store <8 x i16> [[TMP4]], <8 x i16>* bitcast (i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 8) to <8 x i16>*), align 2
; SSE-NEXT: ret void
;
; AVX-LABEL: @bitreverse_16i16(
; AVX-NEXT: [[TMP1:%.*]] = load <16 x i16>, <16 x i16>* bitcast ([16 x i16]* @src16 to <16 x i16>*), align 2
; AVX-NEXT: [[TMP2:%.*]] = call <16 x i16> @llvm.bitreverse.v16i16(<16 x i16> [[TMP1]])
; AVX-NEXT: store <16 x i16> [[TMP2]], <16 x i16>* bitcast ([16 x i16]* @dst16 to <16 x i16>*), align 2
; AVX-NEXT: ret void
;
; XOP-LABEL: @bitreverse_16i16(
; XOP-NEXT: [[TMP1:%.*]] = load <16 x i16>, <16 x i16>* bitcast ([16 x i16]* @src16 to <16 x i16>*), align 2
; XOP-NEXT: [[TMP2:%.*]] = call <16 x i16> @llvm.bitreverse.v16i16(<16 x i16> [[TMP1]])
; XOP-NEXT: store <16 x i16> [[TMP2]], <16 x i16>* bitcast ([16 x i16]* @dst16 to <16 x i16>*), align 2
; XOP-NEXT: ret void
;
%ld0 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 0), align 2
%ld1 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 1), align 2
%ld2 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 2), align 2
%ld3 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 3), align 2
%ld4 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 4), align 2
%ld5 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 5), align 2
%ld6 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 6), align 2
%ld7 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 7), align 2
%ld8 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 8), align 2
%ld9 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 9), align 2
%ld10 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 10), align 2
%ld11 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 11), align 2
%ld12 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 12), align 2
%ld13 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 13), align 2
%ld14 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 14), align 2
%ld15 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 15), align 2
%bitreverse0 = call i16 @llvm.bitreverse.i16(i16 %ld0)
%bitreverse1 = call i16 @llvm.bitreverse.i16(i16 %ld1)
%bitreverse2 = call i16 @llvm.bitreverse.i16(i16 %ld2)
%bitreverse3 = call i16 @llvm.bitreverse.i16(i16 %ld3)
%bitreverse4 = call i16 @llvm.bitreverse.i16(i16 %ld4)
%bitreverse5 = call i16 @llvm.bitreverse.i16(i16 %ld5)
%bitreverse6 = call i16 @llvm.bitreverse.i16(i16 %ld6)
%bitreverse7 = call i16 @llvm.bitreverse.i16(i16 %ld7)
%bitreverse8 = call i16 @llvm.bitreverse.i16(i16 %ld8)
%bitreverse9 = call i16 @llvm.bitreverse.i16(i16 %ld9)
%bitreverse10 = call i16 @llvm.bitreverse.i16(i16 %ld10)
%bitreverse11 = call i16 @llvm.bitreverse.i16(i16 %ld11)
%bitreverse12 = call i16 @llvm.bitreverse.i16(i16 %ld12)
%bitreverse13 = call i16 @llvm.bitreverse.i16(i16 %ld13)
%bitreverse14 = call i16 @llvm.bitreverse.i16(i16 %ld14)
%bitreverse15 = call i16 @llvm.bitreverse.i16(i16 %ld15)
store i16 %bitreverse0 , i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 0), align 2
store i16 %bitreverse1 , i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 1), align 2
store i16 %bitreverse2 , i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 2), align 2
store i16 %bitreverse3 , i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 3), align 2
store i16 %bitreverse4 , i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 4), align 2
store i16 %bitreverse5 , i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 5), align 2
store i16 %bitreverse6 , i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 6), align 2
store i16 %bitreverse7 , i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 7), align 2
store i16 %bitreverse8 , i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 8), align 2
store i16 %bitreverse9 , i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 9), align 2
store i16 %bitreverse10, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 10), align 2
store i16 %bitreverse11, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 11), align 2
store i16 %bitreverse12, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 12), align 2
store i16 %bitreverse13, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 13), align 2
store i16 %bitreverse14, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 14), align 2
store i16 %bitreverse15, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 15), align 2
ret void
}
define void @bitreverse_16i8() #0 {
; CHECK-LABEL: @bitreverse_16i8(
; CHECK-NEXT: [[TMP1:%.*]] = load <16 x i8>, <16 x i8>* bitcast ([32 x i8]* @src8 to <16 x i8>*), align 1
; CHECK-NEXT: [[TMP2:%.*]] = call <16 x i8> @llvm.bitreverse.v16i8(<16 x i8> [[TMP1]])
; CHECK-NEXT: store <16 x i8> [[TMP2]], <16 x i8>* bitcast ([32 x i8]* @dst8 to <16 x i8>*), align 1
; CHECK-NEXT: ret void
;
%ld0 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 0), align 1
%ld1 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 1), align 1
%ld2 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 2), align 1
%ld3 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 3), align 1
%ld4 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 4), align 1
%ld5 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 5), align 1
%ld6 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 6), align 1
%ld7 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 7), align 1
%ld8 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 8), align 1
%ld9 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 9), align 1
%ld10 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 10), align 1
%ld11 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 11), align 1
%ld12 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 12), align 1
%ld13 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 13), align 1
%ld14 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 14), align 1
%ld15 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 15), align 1
%bitreverse0 = call i8 @llvm.bitreverse.i8(i8 %ld0)
%bitreverse1 = call i8 @llvm.bitreverse.i8(i8 %ld1)
%bitreverse2 = call i8 @llvm.bitreverse.i8(i8 %ld2)
%bitreverse3 = call i8 @llvm.bitreverse.i8(i8 %ld3)
%bitreverse4 = call i8 @llvm.bitreverse.i8(i8 %ld4)
%bitreverse5 = call i8 @llvm.bitreverse.i8(i8 %ld5)
%bitreverse6 = call i8 @llvm.bitreverse.i8(i8 %ld6)
%bitreverse7 = call i8 @llvm.bitreverse.i8(i8 %ld7)
%bitreverse8 = call i8 @llvm.bitreverse.i8(i8 %ld8)
%bitreverse9 = call i8 @llvm.bitreverse.i8(i8 %ld9)
%bitreverse10 = call i8 @llvm.bitreverse.i8(i8 %ld10)
%bitreverse11 = call i8 @llvm.bitreverse.i8(i8 %ld11)
%bitreverse12 = call i8 @llvm.bitreverse.i8(i8 %ld12)
%bitreverse13 = call i8 @llvm.bitreverse.i8(i8 %ld13)
%bitreverse14 = call i8 @llvm.bitreverse.i8(i8 %ld14)
%bitreverse15 = call i8 @llvm.bitreverse.i8(i8 %ld15)
store i8 %bitreverse0 , i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 0), align 1
store i8 %bitreverse1 , i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 1), align 1
store i8 %bitreverse2 , i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 2), align 1
store i8 %bitreverse3 , i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 3), align 1
store i8 %bitreverse4 , i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 4), align 1
store i8 %bitreverse5 , i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 5), align 1
store i8 %bitreverse6 , i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 6), align 1
store i8 %bitreverse7 , i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 7), align 1
store i8 %bitreverse8 , i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 8), align 1
store i8 %bitreverse9 , i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 9), align 1
store i8 %bitreverse10, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 10), align 1
store i8 %bitreverse11, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 11), align 1
store i8 %bitreverse12, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 12), align 1
store i8 %bitreverse13, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 13), align 1
store i8 %bitreverse14, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 14), align 1
store i8 %bitreverse15, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 15), align 1
ret void
}
define void @bitreverse_32i8() #0 {
; CHECK-LABEL: @bitreverse_32i8(
; CHECK-NEXT: [[TMP1:%.*]] = load <16 x i8>, <16 x i8>* bitcast ([32 x i8]* @src8 to <16 x i8>*), align 1
; CHECK-NEXT: [[TMP2:%.*]] = load <16 x i8>, <16 x i8>* bitcast (i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 16) to <16 x i8>*), align 1
; CHECK-NEXT: [[TMP3:%.*]] = call <16 x i8> @llvm.bitreverse.v16i8(<16 x i8> [[TMP1]])
; CHECK-NEXT: [[TMP4:%.*]] = call <16 x i8> @llvm.bitreverse.v16i8(<16 x i8> [[TMP2]])
; CHECK-NEXT: store <16 x i8> [[TMP3]], <16 x i8>* bitcast ([32 x i8]* @dst8 to <16 x i8>*), align 1
; CHECK-NEXT: store <16 x i8> [[TMP4]], <16 x i8>* bitcast (i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 16) to <16 x i8>*), align 1
; CHECK-NEXT: ret void
;
%ld0 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 0), align 1
%ld1 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 1), align 1
%ld2 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 2), align 1
%ld3 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 3), align 1
%ld4 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 4), align 1
%ld5 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 5), align 1
%ld6 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 6), align 1
%ld7 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 7), align 1
%ld8 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 8), align 1
%ld9 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 9), align 1
%ld10 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 10), align 1
%ld11 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 11), align 1
%ld12 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 12), align 1
%ld13 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 13), align 1
%ld14 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 14), align 1
%ld15 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 15), align 1
%ld16 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 16), align 1
%ld17 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 17), align 1
%ld18 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 18), align 1
%ld19 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 19), align 1
%ld20 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 20), align 1
%ld21 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 21), align 1
%ld22 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 22), align 1
%ld23 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 23), align 1
%ld24 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 24), align 1
%ld25 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 25), align 1
%ld26 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 26), align 1
%ld27 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 27), align 1
%ld28 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 28), align 1
%ld29 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 29), align 1
%ld30 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 30), align 1
%ld31 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 31), align 1
%bitreverse0 = call i8 @llvm.bitreverse.i8(i8 %ld0)
%bitreverse1 = call i8 @llvm.bitreverse.i8(i8 %ld1)
%bitreverse2 = call i8 @llvm.bitreverse.i8(i8 %ld2)
%bitreverse3 = call i8 @llvm.bitreverse.i8(i8 %ld3)
%bitreverse4 = call i8 @llvm.bitreverse.i8(i8 %ld4)
%bitreverse5 = call i8 @llvm.bitreverse.i8(i8 %ld5)
%bitreverse6 = call i8 @llvm.bitreverse.i8(i8 %ld6)
%bitreverse7 = call i8 @llvm.bitreverse.i8(i8 %ld7)
%bitreverse8 = call i8 @llvm.bitreverse.i8(i8 %ld8)
%bitreverse9 = call i8 @llvm.bitreverse.i8(i8 %ld9)
%bitreverse10 = call i8 @llvm.bitreverse.i8(i8 %ld10)
%bitreverse11 = call i8 @llvm.bitreverse.i8(i8 %ld11)
%bitreverse12 = call i8 @llvm.bitreverse.i8(i8 %ld12)
%bitreverse13 = call i8 @llvm.bitreverse.i8(i8 %ld13)
%bitreverse14 = call i8 @llvm.bitreverse.i8(i8 %ld14)
%bitreverse15 = call i8 @llvm.bitreverse.i8(i8 %ld15)
%bitreverse16 = call i8 @llvm.bitreverse.i8(i8 %ld16)
%bitreverse17 = call i8 @llvm.bitreverse.i8(i8 %ld17)
%bitreverse18 = call i8 @llvm.bitreverse.i8(i8 %ld18)
%bitreverse19 = call i8 @llvm.bitreverse.i8(i8 %ld19)
%bitreverse20 = call i8 @llvm.bitreverse.i8(i8 %ld20)
%bitreverse21 = call i8 @llvm.bitreverse.i8(i8 %ld21)
%bitreverse22 = call i8 @llvm.bitreverse.i8(i8 %ld22)
%bitreverse23 = call i8 @llvm.bitreverse.i8(i8 %ld23)
%bitreverse24 = call i8 @llvm.bitreverse.i8(i8 %ld24)
%bitreverse25 = call i8 @llvm.bitreverse.i8(i8 %ld25)
%bitreverse26 = call i8 @llvm.bitreverse.i8(i8 %ld26)
%bitreverse27 = call i8 @llvm.bitreverse.i8(i8 %ld27)
%bitreverse28 = call i8 @llvm.bitreverse.i8(i8 %ld28)
%bitreverse29 = call i8 @llvm.bitreverse.i8(i8 %ld29)
%bitreverse30 = call i8 @llvm.bitreverse.i8(i8 %ld30)
%bitreverse31 = call i8 @llvm.bitreverse.i8(i8 %ld31)
store i8 %bitreverse0 , i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 0), align 1
store i8 %bitreverse1 , i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 1), align 1
store i8 %bitreverse2 , i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 2), align 1
store i8 %bitreverse3 , i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 3), align 1
store i8 %bitreverse4 , i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 4), align 1
store i8 %bitreverse5 , i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 5), align 1
store i8 %bitreverse6 , i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 6), align 1
store i8 %bitreverse7 , i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 7), align 1
store i8 %bitreverse8 , i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 8), align 1
store i8 %bitreverse9 , i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 9), align 1
store i8 %bitreverse10, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 10), align 1
store i8 %bitreverse11, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 11), align 1
store i8 %bitreverse12, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 12), align 1
store i8 %bitreverse13, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 13), align 1
store i8 %bitreverse14, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 14), align 1
store i8 %bitreverse15, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 15), align 1
store i8 %bitreverse16, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 16), align 1
store i8 %bitreverse17, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 17), align 1
store i8 %bitreverse18, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 18), align 1
store i8 %bitreverse19, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 19), align 1
store i8 %bitreverse20, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 20), align 1
store i8 %bitreverse21, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 21), align 1
store i8 %bitreverse22, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 22), align 1
store i8 %bitreverse23, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 23), align 1
store i8 %bitreverse24, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 24), align 1
store i8 %bitreverse25, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 25), align 1
store i8 %bitreverse26, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 26), align 1
store i8 %bitreverse27, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 27), align 1
store i8 %bitreverse28, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 28), align 1
store i8 %bitreverse29, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 29), align 1
store i8 %bitreverse30, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 30), align 1
store i8 %bitreverse31, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 31), align 1
ret void
}
attributes #0 = { nounwind }

166
external/llvm/test/Transforms/SLPVectorizer/X86/blending-shuffle.ll vendored
View File

@ -1,166 +0,0 @@
; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
; RUN: opt < %s -slp-vectorizer -S -o - -mtriple=x86-64-unknown-linux -mcpu=bdver2 -instcombine | FileCheck %s
define <2 x i8> @g(<2 x i8> %x, <2 x i8> %y) {
; CHECK-LABEL: @g(
; CHECK-NEXT: [[TMP1:%.*]] = shufflevector <2 x i8> [[X:%.*]], <2 x i8> [[Y:%.*]], <2 x i32> <i32 0, i32 3>
; CHECK-NEXT: [[TMP2:%.*]] = mul <2 x i8> [[TMP1]], [[TMP1]]
; CHECK-NEXT: ret <2 x i8> [[TMP2]]
;
%x0 = extractelement <2 x i8> %x, i32 0
%y1 = extractelement <2 x i8> %y, i32 1
%x0x0 = mul i8 %x0, %x0
%y1y1 = mul i8 %y1, %y1
%ins1 = insertelement <2 x i8> undef, i8 %x0x0, i32 0
%ins2 = insertelement <2 x i8> %ins1, i8 %y1y1, i32 1
ret <2 x i8> %ins2
}
define <4 x i8> @h(<4 x i8> %x, <4 x i8> %y) {
; CHECK-LABEL: @h(
; CHECK-NEXT: [[TMP1:%.*]] = shufflevector <4 x i8> [[X:%.*]], <4 x i8> [[Y:%.*]], <4 x i32> <i32 0, i32 3, i32 5, i32 6>
; CHECK-NEXT: [[TMP2:%.*]] = mul <4 x i8> [[TMP1]], [[TMP1]]
; CHECK-NEXT: ret <4 x i8> [[TMP2]]
;
%x0 = extractelement <4 x i8> %x, i32 0
%x3 = extractelement <4 x i8> %x, i32 3
%y1 = extractelement <4 x i8> %y, i32 1
%y2 = extractelement <4 x i8> %y, i32 2
%x0x0 = mul i8 %x0, %x0
%x3x3 = mul i8 %x3, %x3
%y1y1 = mul i8 %y1, %y1
%y2y2 = mul i8 %y2, %y2
%ins1 = insertelement <4 x i8> undef, i8 %x0x0, i32 0
%ins2 = insertelement <4 x i8> %ins1, i8 %x3x3, i32 1
%ins3 = insertelement <4 x i8> %ins2, i8 %y1y1, i32 2
%ins4 = insertelement <4 x i8> %ins3, i8 %y2y2, i32 3
ret <4 x i8> %ins4
}
define <4 x i8> @h_undef(<4 x i8> %x, <4 x i8> %y) {
; CHECK-LABEL: @h_undef(
; CHECK-NEXT: [[TMP1:%.*]] = shufflevector <4 x i8> [[X:%.*]], <4 x i8> [[Y:%.*]], <4 x i32> <i32 undef, i32 3, i32 5, i32 6>
; CHECK-NEXT: [[TMP2:%.*]] = mul <4 x i8> [[TMP1]], [[TMP1]]
; CHECK-NEXT: ret <4 x i8> [[TMP2]]
;
%x0 = extractelement <4 x i8> undef, i32 0
%x3 = extractelement <4 x i8> %x, i32 3
%y1 = extractelement <4 x i8> %y, i32 1
%y2 = extractelement <4 x i8> %y, i32 2
%x0x0 = mul i8 %x0, %x0
%x3x3 = mul i8 %x3, %x3
%y1y1 = mul i8 %y1, %y1
%y2y2 = mul i8 %y2, %y2
%ins1 = insertelement <4 x i8> undef, i8 %x0x0, i32 0
%ins2 = insertelement <4 x i8> %ins1, i8 %x3x3, i32 1
%ins3 = insertelement <4 x i8> %ins2, i8 %y1y1, i32 2
%ins4 = insertelement <4 x i8> %ins3, i8 %y2y2, i32 3
ret <4 x i8> %ins4
}
define i8 @i(<4 x i8> %x, <4 x i8> %y) {
; CHECK-LABEL: @i(
; CHECK-NEXT: [[TMP1:%.*]] = shufflevector <4 x i8> [[X:%.*]], <4 x i8> [[Y:%.*]], <4 x i32> <i32 0, i32 3, i32 5, i32 6>
; CHECK-NEXT: [[TMP2:%.*]] = mul <4 x i8> [[TMP1]], [[TMP1]]
; CHECK-NEXT: [[RDX_SHUF:%.*]] = shufflevector <4 x i8> [[TMP2]], <4 x i8> undef, <4 x i32> <i32 2, i32 3, i32 undef, i32 undef>
; CHECK-NEXT: [[BIN_RDX:%.*]] = add <4 x i8> [[TMP2]], [[RDX_SHUF]]
; CHECK-NEXT: [[RDX_SHUF1:%.*]] = shufflevector <4 x i8> [[BIN_RDX]], <4 x i8> undef, <4 x i32> <i32 1, i32 undef, i32 undef, i32 undef>
; CHECK-NEXT: [[BIN_RDX2:%.*]] = add <4 x i8> [[BIN_RDX]], [[RDX_SHUF1]]
; CHECK-NEXT: [[TMP3:%.*]] = extractelement <4 x i8> [[BIN_RDX2]], i32 0
; CHECK-NEXT: ret i8 [[TMP3]]
;
%x0 = extractelement <4 x i8> %x, i32 0
%x3 = extractelement <4 x i8> %x, i32 3
%y1 = extractelement <4 x i8> %y, i32 1
%y2 = extractelement <4 x i8> %y, i32 2
%x0x0 = mul i8 %x0, %x0
%x3x3 = mul i8 %x3, %x3
%y1y1 = mul i8 %y1, %y1
%y2y2 = mul i8 %y2, %y2
%1 = add i8 %x0x0, %x3x3
%2 = add i8 %y1y1, %y2y2
%3 = add i8 %1, %2
ret i8 %3
}
define i8 @j(<4 x i8> %x, <4 x i8> %y) {
; CHECK-LABEL: @j(
; CHECK-NEXT: [[TMP1:%.*]] = shufflevector <4 x i8> [[X:%.*]], <4 x i8> [[Y:%.*]], <2 x i32> <i32 0, i32 5>
; CHECK-NEXT: [[TMP2:%.*]] = mul <2 x i8> [[TMP1]], [[TMP1]]
; CHECK-NEXT: [[TMP3:%.*]] = shufflevector <4 x i8> [[X]], <4 x i8> [[Y]], <2 x i32> <i32 3, i32 6>
; CHECK-NEXT: [[TMP4:%.*]] = mul <2 x i8> [[TMP3]], [[TMP3]]
; CHECK-NEXT: [[TMP5:%.*]] = add <2 x i8> [[TMP2]], [[TMP4]]
; CHECK-NEXT: [[TMP6:%.*]] = extractelement <2 x i8> [[TMP5]], i32 0
; CHECK-NEXT: [[TMP7:%.*]] = extractelement <2 x i8> [[TMP5]], i32 1
; CHECK-NEXT: [[TMP8:%.*]] = sdiv i8 [[TMP6]], [[TMP7]]
; CHECK-NEXT: ret i8 [[TMP8]]
;
%x0 = extractelement <4 x i8> %x, i32 0
%x3 = extractelement <4 x i8> %x, i32 3
%y1 = extractelement <4 x i8> %y, i32 1
%y2 = extractelement <4 x i8> %y, i32 2
%x0x0 = mul i8 %x0, %x0
%x3x3 = mul i8 %x3, %x3
%y1y1 = mul i8 %y1, %y1
%y2y2 = mul i8 %y2, %y2
%1 = add i8 %x0x0, %x3x3
%2 = add i8 %y1y1, %y2y2
%3 = sdiv i8 %1, %2
ret i8 %3
}
define i8 @k(<4 x i8> %x) {
; CHECK-LABEL: @k(
; CHECK-NEXT: [[TMP1:%.*]] = mul <4 x i8> [[X:%.*]], [[X]]
; CHECK-NEXT: [[TMP2:%.*]] = shufflevector <4 x i8> [[TMP1]], <4 x i8> undef, <2 x i32> <i32 0, i32 1>
; CHECK-NEXT: [[TMP3:%.*]] = mul <4 x i8> [[X]], [[X]]
; CHECK-NEXT: [[TMP4:%.*]] = shufflevector <4 x i8> [[TMP3]], <4 x i8> undef, <2 x i32> <i32 3, i32 2>
; CHECK-NEXT: [[TMP5:%.*]] = add <2 x i8> [[TMP2]], [[TMP4]]
; CHECK-NEXT: [[TMP6:%.*]] = extractelement <2 x i8> [[TMP5]], i32 0
; CHECK-NEXT: [[TMP7:%.*]] = extractelement <2 x i8> [[TMP5]], i32 1
; CHECK-NEXT: [[TMP8:%.*]] = sdiv i8 [[TMP6]], [[TMP7]]
; CHECK-NEXT: ret i8 [[TMP8]]
;
%x0 = extractelement <4 x i8> %x, i32 0
%x3 = extractelement <4 x i8> %x, i32 3
%x1 = extractelement <4 x i8> %x, i32 1
%x2 = extractelement <4 x i8> %x, i32 2
%x0x0 = mul i8 %x0, %x0
%x3x3 = mul i8 %x3, %x3
%x1x1 = mul i8 %x1, %x1
%x2x2 = mul i8 %x2, %x2
%1 = add i8 %x0x0, %x3x3
%2 = add i8 %x1x1, %x2x2
%3 = sdiv i8 %1, %2
ret i8 %3
}
define i8 @k_bb(<4 x i8> %x) {
; CHECK-LABEL: @k_bb(
; CHECK-NEXT: br label [[BB1:%.*]]
; CHECK: bb1:
; CHECK-NEXT: [[TMP1:%.*]] = mul <4 x i8> [[X:%.*]], [[X]]
; CHECK-NEXT: [[TMP2:%.*]] = shufflevector <4 x i8> [[TMP1]], <4 x i8> undef, <2 x i32> <i32 0, i32 1>
; CHECK-NEXT: [[TMP3:%.*]] = mul <4 x i8> [[X]], [[X]]
; CHECK-NEXT: [[TMP4:%.*]] = shufflevector <4 x i8> [[TMP3]], <4 x i8> undef, <2 x i32> <i32 3, i32 2>
; CHECK-NEXT: [[TMP5:%.*]] = add <2 x i8> [[TMP2]], [[TMP4]]
; CHECK-NEXT: [[TMP6:%.*]] = extractelement <2 x i8> [[TMP5]], i32 0
; CHECK-NEXT: [[TMP7:%.*]] = extractelement <2 x i8> [[TMP5]], i32 1
; CHECK-NEXT: [[TMP8:%.*]] = sdiv i8 [[TMP6]], [[TMP7]]
; CHECK-NEXT: ret i8 [[TMP8]]
;
%x0 = extractelement <4 x i8> %x, i32 0
br label %bb1
bb1:
%x3 = extractelement <4 x i8> %x, i32 3
%x1 = extractelement <4 x i8> %x, i32 1
%x2 = extractelement <4 x i8> %x, i32 2
%x0x0 = mul i8 %x0, %x0
%x3x3 = mul i8 %x3, %x3
%x1x1 = mul i8 %x1, %x1
%x2x2 = mul i8 %x2, %x2
%1 = add i8 %x0x0, %x3x3
%2 = add i8 %x1x1, %x2x2
%3 = sdiv i8 %1, %2
ret i8 %3
}

247
external/llvm/test/Transforms/SLPVectorizer/X86/bswap.ll vendored
View File

@ -1,247 +0,0 @@
; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
; RUN: opt < %s -mtriple=x86_64-unknown -basicaa -slp-vectorizer -S | FileCheck %s --check-prefix=CHECK --check-prefix=SSE
; RUN: opt < %s -mtriple=x86_64-unknown -mcpu=corei7-avx -basicaa -slp-vectorizer -S | FileCheck %s --check-prefix=CHECK --check-prefix=AVX
; RUN: opt < %s -mtriple=x86_64-unknown -mcpu=core-avx2 -basicaa -slp-vectorizer -S | FileCheck %s --check-prefix=CHECK --check-prefix=AVX
target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
@src64 = common global [4 x i64] zeroinitializer, align 32
@dst64 = common global [4 x i64] zeroinitializer, align 32
@src32 = common global [8 x i32] zeroinitializer, align 32
@dst32 = common global [8 x i32] zeroinitializer, align 32
@src16 = common global [16 x i16] zeroinitializer, align 32
@dst16 = common global [16 x i16] zeroinitializer, align 32
declare i64 @llvm.bswap.i64(i64)
declare i32 @llvm.bswap.i32(i32)
declare i16 @llvm.bswap.i16(i16)
define void @bswap_2i64() #0 {
; SSE-LABEL: @bswap_2i64(
; SSE-NEXT: [[LD0:%.*]] = load i64, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @src64, i32 0, i64 0), align 8
; SSE-NEXT: [[LD1:%.*]] = load i64, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @src64, i32 0, i64 1), align 8
; SSE-NEXT: [[BSWAP0:%.*]] = call i64 @llvm.bswap.i64(i64 [[LD0]])
; SSE-NEXT: [[BSWAP1:%.*]] = call i64 @llvm.bswap.i64(i64 [[LD1]])
; SSE-NEXT: store i64 [[BSWAP0]], i64* getelementptr inbounds ([4 x i64], [4 x i64]* @dst64, i32 0, i64 0), align 8
; SSE-NEXT: store i64 [[BSWAP1]], i64* getelementptr inbounds ([4 x i64], [4 x i64]* @dst64, i32 0, i64 1), align 8
; SSE-NEXT: ret void
;
; AVX-LABEL: @bswap_2i64(
; AVX-NEXT: [[TMP1:%.*]] = load <2 x i64>, <2 x i64>* bitcast ([4 x i64]* @src64 to <2 x i64>*), align 8
; AVX-NEXT: [[TMP2:%.*]] = call <2 x i64> @llvm.bswap.v2i64(<2 x i64> [[TMP1]])
; AVX-NEXT: store <2 x i64> [[TMP2]], <2 x i64>* bitcast ([4 x i64]* @dst64 to <2 x i64>*), align 8
; AVX-NEXT: ret void
;
%ld0 = load i64, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @src64, i32 0, i64 0), align 8
%ld1 = load i64, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @src64, i32 0, i64 1), align 8
%bswap0 = call i64 @llvm.bswap.i64(i64 %ld0)
%bswap1 = call i64 @llvm.bswap.i64(i64 %ld1)
store i64 %bswap0, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @dst64, i32 0, i64 0), align 8
store i64 %bswap1, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @dst64, i32 0, i64 1), align 8
ret void
}
define void @bswap_4i64() #0 {
; SSE-LABEL: @bswap_4i64(
; SSE-NEXT: [[LD0:%.*]] = load i64, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @src64, i64 0, i64 0), align 4
; SSE-NEXT: [[LD1:%.*]] = load i64, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @src64, i64 0, i64 1), align 4
; SSE-NEXT: [[LD2:%.*]] = load i64, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @src64, i64 0, i64 2), align 4
; SSE-NEXT: [[LD3:%.*]] = load i64, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @src64, i64 0, i64 3), align 4
; SSE-NEXT: [[BSWAP0:%.*]] = call i64 @llvm.bswap.i64(i64 [[LD0]])
; SSE-NEXT: [[BSWAP1:%.*]] = call i64 @llvm.bswap.i64(i64 [[LD1]])
; SSE-NEXT: [[BSWAP2:%.*]] = call i64 @llvm.bswap.i64(i64 [[LD2]])
; SSE-NEXT: [[BSWAP3:%.*]] = call i64 @llvm.bswap.i64(i64 [[LD3]])
; SSE-NEXT: store i64 [[BSWAP0]], i64* getelementptr inbounds ([4 x i64], [4 x i64]* @dst64, i64 0, i64 0), align 4
; SSE-NEXT: store i64 [[BSWAP1]], i64* getelementptr inbounds ([4 x i64], [4 x i64]* @dst64, i64 0, i64 1), align 4
; SSE-NEXT: store i64 [[BSWAP2]], i64* getelementptr inbounds ([4 x i64], [4 x i64]* @dst64, i64 0, i64 2), align 4
; SSE-NEXT: store i64 [[BSWAP3]], i64* getelementptr inbounds ([4 x i64], [4 x i64]* @dst64, i64 0, i64 3), align 4
; SSE-NEXT: ret void
;
; AVX-LABEL: @bswap_4i64(
; AVX-NEXT: [[TMP1:%.*]] = load <4 x i64>, <4 x i64>* bitcast ([4 x i64]* @src64 to <4 x i64>*), align 4
; AVX-NEXT: [[TMP2:%.*]] = call <4 x i64> @llvm.bswap.v4i64(<4 x i64> [[TMP1]])
; AVX-NEXT: store <4 x i64> [[TMP2]], <4 x i64>* bitcast ([4 x i64]* @dst64 to <4 x i64>*), align 4
; AVX-NEXT: ret void
;
%ld0 = load i64, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @src64, i64 0, i64 0), align 4
%ld1 = load i64, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @src64, i64 0, i64 1), align 4
%ld2 = load i64, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @src64, i64 0, i64 2), align 4
%ld3 = load i64, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @src64, i64 0, i64 3), align 4
%bswap0 = call i64 @llvm.bswap.i64(i64 %ld0)
%bswap1 = call i64 @llvm.bswap.i64(i64 %ld1)
%bswap2 = call i64 @llvm.bswap.i64(i64 %ld2)
%bswap3 = call i64 @llvm.bswap.i64(i64 %ld3)
store i64 %bswap0, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @dst64, i64 0, i64 0), align 4
store i64 %bswap1, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @dst64, i64 0, i64 1), align 4
store i64 %bswap2, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @dst64, i64 0, i64 2), align 4
store i64 %bswap3, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @dst64, i64 0, i64 3), align 4
ret void
}
define void @bswap_4i32() #0 {
; CHECK-LABEL: @bswap_4i32(
; CHECK-NEXT: [[TMP1:%.*]] = load <4 x i32>, <4 x i32>* bitcast ([8 x i32]* @src32 to <4 x i32>*), align 4
; CHECK-NEXT: [[TMP2:%.*]] = call <4 x i32> @llvm.bswap.v4i32(<4 x i32> [[TMP1]])
; CHECK-NEXT: store <4 x i32> [[TMP2]], <4 x i32>* bitcast ([8 x i32]* @dst32 to <4 x i32>*), align 4
; CHECK-NEXT: ret void
;
%ld0 = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 0), align 4
%ld1 = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 1), align 4
%ld2 = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 2), align 4
%ld3 = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 3), align 4
%bswap0 = call i32 @llvm.bswap.i32(i32 %ld0)
%bswap1 = call i32 @llvm.bswap.i32(i32 %ld1)
%bswap2 = call i32 @llvm.bswap.i32(i32 %ld2)
%bswap3 = call i32 @llvm.bswap.i32(i32 %ld3)
store i32 %bswap0, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 0), align 4
store i32 %bswap1, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 1), align 4
store i32 %bswap2, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 2), align 4
store i32 %bswap3, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 3), align 4
ret void
}
define void @bswap_8i32() #0 {
; SSE-LABEL: @bswap_8i32(
; SSE-NEXT: [[TMP1:%.*]] = load <4 x i32>, <4 x i32>* bitcast ([8 x i32]* @src32 to <4 x i32>*), align 2
; SSE-NEXT: [[TMP2:%.*]] = load <4 x i32>, <4 x i32>* bitcast (i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 4) to <4 x i32>*), align 2
; SSE-NEXT: [[TMP3:%.*]] = call <4 x i32> @llvm.bswap.v4i32(<4 x i32> [[TMP1]])
; SSE-NEXT: [[TMP4:%.*]] = call <4 x i32> @llvm.bswap.v4i32(<4 x i32> [[TMP2]])
; SSE-NEXT: store <4 x i32> [[TMP3]], <4 x i32>* bitcast ([8 x i32]* @dst32 to <4 x i32>*), align 2
; SSE-NEXT: store <4 x i32> [[TMP4]], <4 x i32>* bitcast (i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 4) to <4 x i32>*), align 2
; SSE-NEXT: ret void
;
; AVX-LABEL: @bswap_8i32(
; AVX-NEXT: [[TMP1:%.*]] = load <8 x i32>, <8 x i32>* bitcast ([8 x i32]* @src32 to <8 x i32>*), align 2
; AVX-NEXT: [[TMP2:%.*]] = call <8 x i32> @llvm.bswap.v8i32(<8 x i32> [[TMP1]])
; AVX-NEXT: store <8 x i32> [[TMP2]], <8 x i32>* bitcast ([8 x i32]* @dst32 to <8 x i32>*), align 2
; AVX-NEXT: ret void
;
%ld0 = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 0), align 2
%ld1 = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 1), align 2
%ld2 = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 2), align 2
%ld3 = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 3), align 2
%ld4 = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 4), align 2
%ld5 = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 5), align 2
%ld6 = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 6), align 2
%ld7 = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 7), align 2
%bswap0 = call i32 @llvm.bswap.i32(i32 %ld0)
%bswap1 = call i32 @llvm.bswap.i32(i32 %ld1)
%bswap2 = call i32 @llvm.bswap.i32(i32 %ld2)
%bswap3 = call i32 @llvm.bswap.i32(i32 %ld3)
%bswap4 = call i32 @llvm.bswap.i32(i32 %ld4)
%bswap5 = call i32 @llvm.bswap.i32(i32 %ld5)
%bswap6 = call i32 @llvm.bswap.i32(i32 %ld6)
%bswap7 = call i32 @llvm.bswap.i32(i32 %ld7)
store i32 %bswap0, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 0), align 2
store i32 %bswap1, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 1), align 2
store i32 %bswap2, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 2), align 2
store i32 %bswap3, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 3), align 2
store i32 %bswap4, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 4), align 2
store i32 %bswap5, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 5), align 2
store i32 %bswap6, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 6), align 2
store i32 %bswap7, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 7), align 2
ret void
}
define void @bswap_8i16() #0 {
; CHECK-LABEL: @bswap_8i16(
; CHECK-NEXT: [[TMP1:%.*]] = load <8 x i16>, <8 x i16>* bitcast ([16 x i16]* @src16 to <8 x i16>*), align 2
; CHECK-NEXT: [[TMP2:%.*]] = call <8 x i16> @llvm.bswap.v8i16(<8 x i16> [[TMP1]])
; CHECK-NEXT: store <8 x i16> [[TMP2]], <8 x i16>* bitcast ([16 x i16]* @dst16 to <8 x i16>*), align 2
; CHECK-NEXT: ret void
;
%ld0 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 0), align 2
%ld1 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 1), align 2
%ld2 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 2), align 2
%ld3 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 3), align 2
%ld4 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 4), align 2
%ld5 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 5), align 2
%ld6 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 6), align 2
%ld7 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 7), align 2
%bswap0 = call i16 @llvm.bswap.i16(i16 %ld0)
%bswap1 = call i16 @llvm.bswap.i16(i16 %ld1)
%bswap2 = call i16 @llvm.bswap.i16(i16 %ld2)
%bswap3 = call i16 @llvm.bswap.i16(i16 %ld3)
%bswap4 = call i16 @llvm.bswap.i16(i16 %ld4)
%bswap5 = call i16 @llvm.bswap.i16(i16 %ld5)
%bswap6 = call i16 @llvm.bswap.i16(i16 %ld6)
%bswap7 = call i16 @llvm.bswap.i16(i16 %ld7)
store i16 %bswap0, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 0), align 2
store i16 %bswap1, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 1), align 2
store i16 %bswap2, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 2), align 2
store i16 %bswap3, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 3), align 2
store i16 %bswap4, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 4), align 2
store i16 %bswap5, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 5), align 2
store i16 %bswap6, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 6), align 2
store i16 %bswap7, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 7), align 2
ret void
}
define void @bswap_16i16() #0 {
; SSE-LABEL: @bswap_16i16(
; SSE-NEXT: [[TMP1:%.*]] = load <8 x i16>, <8 x i16>* bitcast ([16 x i16]* @src16 to <8 x i16>*), align 2
; SSE-NEXT: [[TMP2:%.*]] = load <8 x i16>, <8 x i16>* bitcast (i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 8) to <8 x i16>*), align 2
; SSE-NEXT: [[TMP3:%.*]] = call <8 x i16> @llvm.bswap.v8i16(<8 x i16> [[TMP1]])
; SSE-NEXT: [[TMP4:%.*]] = call <8 x i16> @llvm.bswap.v8i16(<8 x i16> [[TMP2]])
; SSE-NEXT: store <8 x i16> [[TMP3]], <8 x i16>* bitcast ([16 x i16]* @dst16 to <8 x i16>*), align 2
; SSE-NEXT: store <8 x i16> [[TMP4]], <8 x i16>* bitcast (i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 8) to <8 x i16>*), align 2
; SSE-NEXT: ret void
;
; AVX-LABEL: @bswap_16i16(
; AVX-NEXT: [[TMP1:%.*]] = load <16 x i16>, <16 x i16>* bitcast ([16 x i16]* @src16 to <16 x i16>*), align 2
; AVX-NEXT: [[TMP2:%.*]] = call <16 x i16> @llvm.bswap.v16i16(<16 x i16> [[TMP1]])
; AVX-NEXT: store <16 x i16> [[TMP2]], <16 x i16>* bitcast ([16 x i16]* @dst16 to <16 x i16>*), align 2
; AVX-NEXT: ret void
;
%ld0 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 0), align 2
%ld1 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 1), align 2
%ld2 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 2), align 2
%ld3 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 3), align 2
%ld4 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 4), align 2
%ld5 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 5), align 2
%ld6 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 6), align 2
%ld7 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 7), align 2
%ld8 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 8), align 2
%ld9 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 9), align 2
%ld10 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 10), align 2
%ld11 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 11), align 2
%ld12 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 12), align 2
%ld13 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 13), align 2
%ld14 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 14), align 2
%ld15 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 15), align 2
%bswap0 = call i16 @llvm.bswap.i16(i16 %ld0)
%bswap1 = call i16 @llvm.bswap.i16(i16 %ld1)
%bswap2 = call i16 @llvm.bswap.i16(i16 %ld2)
%bswap3 = call i16 @llvm.bswap.i16(i16 %ld3)
%bswap4 = call i16 @llvm.bswap.i16(i16 %ld4)
%bswap5 = call i16 @llvm.bswap.i16(i16 %ld5)
%bswap6 = call i16 @llvm.bswap.i16(i16 %ld6)
%bswap7 = call i16 @llvm.bswap.i16(i16 %ld7)
%bswap8 = call i16 @llvm.bswap.i16(i16 %ld8)
%bswap9 = call i16 @llvm.bswap.i16(i16 %ld9)
%bswap10 = call i16 @llvm.bswap.i16(i16 %ld10)
%bswap11 = call i16 @llvm.bswap.i16(i16 %ld11)
%bswap12 = call i16 @llvm.bswap.i16(i16 %ld12)
%bswap13 = call i16 @llvm.bswap.i16(i16 %ld13)
%bswap14 = call i16 @llvm.bswap.i16(i16 %ld14)
%bswap15 = call i16 @llvm.bswap.i16(i16 %ld15)
store i16 %bswap0 , i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 0), align 2
store i16 %bswap1 , i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 1), align 2
store i16 %bswap2 , i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 2), align 2
store i16 %bswap3 , i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 3), align 2
store i16 %bswap4 , i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 4), align 2
store i16 %bswap5 , i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 5), align 2
store i16 %bswap6 , i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 6), align 2
store i16 %bswap7 , i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 7), align 2
store i16 %bswap8 , i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 8), align 2
store i16 %bswap9 , i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 9), align 2
store i16 %bswap10, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 10), align 2
store i16 %bswap11, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 11), align 2
store i16 %bswap12, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 12), align 2
store i16 %bswap13, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 13), align 2
store i16 %bswap14, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 14), align 2
store i16 %bswap15, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 15), align 2
ret void
}
attributes #0 = { nounwind }

176
external/llvm/test/Transforms/SLPVectorizer/X86/call.ll vendored
View File

@ -1,176 +0,0 @@
; RUN: opt < %s -basicaa -slp-vectorizer -slp-threshold=-999 -dce -S -mtriple=x86_64-apple-macosx10.8.0 -mcpu=corei7-avx | FileCheck %s
target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64-S128"
target triple = "x86_64-apple-macosx10.8.0"
declare double @sin(double)
declare double @cos(double)
declare double @pow(double, double)
declare double @exp2(double)
declare double @sqrt(double)
declare i64 @round(i64)
define void @sin_libm(double* %a, double* %b) {
; CHECK-LABEL: @sin_libm(
; CHECK-NEXT: [[TMP1:%.*]] = bitcast double* %a to <2 x double>*
; CHECK-NEXT: [[TMP2:%.*]] = load <2 x double>, <2 x double>* [[TMP1]], align 8
; CHECK-NEXT: [[TMP3:%.*]] = call <2 x double> @llvm.sin.v2f64(<2 x double> [[TMP2]])
; CHECK-NEXT: [[TMP4:%.*]] = bitcast double* %b to <2 x double>*
; CHECK-NEXT: store <2 x double> [[TMP3]], <2 x double>* [[TMP4]], align 8
; CHECK-NEXT: ret void
;
%a0 = load double, double* %a, align 8
%idx1 = getelementptr inbounds double, double* %a, i64 1
%a1 = load double, double* %idx1, align 8
%sin1 = tail call double @sin(double %a0) nounwind readnone
%sin2 = tail call double @sin(double %a1) nounwind readnone
store double %sin1, double* %b, align 8
%idx2 = getelementptr inbounds double, double* %b, i64 1
store double %sin2, double* %idx2, align 8
ret void
}
define void @cos_libm(double* %a, double* %b) {
; CHECK-LABEL: @cos_libm(
; CHECK-NEXT: [[TMP1:%.*]] = bitcast double* %a to <2 x double>*
; CHECK-NEXT: [[TMP2:%.*]] = load <2 x double>, <2 x double>* [[TMP1]], align 8
; CHECK-NEXT: [[TMP3:%.*]] = call <2 x double> @llvm.cos.v2f64(<2 x double> [[TMP2]])
; CHECK-NEXT: [[TMP4:%.*]] = bitcast double* %b to <2 x double>*
; CHECK-NEXT: store <2 x double> [[TMP3]], <2 x double>* [[TMP4]], align 8
; CHECK-NEXT: ret void
;
%a0 = load double, double* %a, align 8
%idx1 = getelementptr inbounds double, double* %a, i64 1
%a1 = load double, double* %idx1, align 8
%cos1 = tail call double @cos(double %a0) nounwind readnone
%cos2 = tail call double @cos(double %a1) nounwind readnone
store double %cos1, double* %b, align 8
%idx2 = getelementptr inbounds double, double* %b, i64 1
store double %cos2, double* %idx2, align 8
ret void
}
define void @pow_libm(double* %a, double* %b) {
; CHECK-LABEL: @pow_libm(
; CHECK-NEXT: [[TMP1:%.*]] = bitcast double* %a to <2 x double>*
; CHECK-NEXT: [[TMP2:%.*]] = load <2 x double>, <2 x double>* [[TMP1]], align 8
; CHECK-NEXT: [[TMP3:%.*]] = call <2 x double> @llvm.pow.v2f64(<2 x double> [[TMP2]], <2 x double> [[TMP2]])
; CHECK-NEXT: [[TMP4:%.*]] = bitcast double* %b to <2 x double>*
; CHECK-NEXT: store <2 x double> [[TMP3]], <2 x double>* [[TMP4]], align 8
; CHECK-NEXT: ret void
;
%a0 = load double, double* %a, align 8
%idx1 = getelementptr inbounds double, double* %a, i64 1
%a1 = load double, double* %idx1, align 8
%pow1 = tail call double @pow(double %a0, double %a0) nounwind readnone
%pow2 = tail call double @pow(double %a1, double %a1) nounwind readnone
store double %pow1, double* %b, align 8
%idx2 = getelementptr inbounds double, double* %b, i64 1
store double %pow2, double* %idx2, align 8
ret void
}
define void @exp_libm(double* %a, double* %b) {
; CHECK-LABEL: @exp_libm(
; CHECK-NEXT: [[TMP1:%.*]] = bitcast double* %a to <2 x double>*
; CHECK-NEXT: [[TMP2:%.*]] = load <2 x double>, <2 x double>* [[TMP1]], align 8
; CHECK-NEXT: [[TMP3:%.*]] = call <2 x double> @llvm.exp2.v2f64(<2 x double> [[TMP2]])
; CHECK-NEXT: [[TMP4:%.*]] = bitcast double* %b to <2 x double>*
; CHECK-NEXT: store <2 x double> [[TMP3]], <2 x double>* [[TMP4]], align 8
; CHECK-NEXT: ret void
;
%a0 = load double, double* %a, align 8
%idx1 = getelementptr inbounds double, double* %a, i64 1
%a1 = load double, double* %idx1, align 8
%exp1 = tail call double @exp2(double %a0) nounwind readnone
%exp2 = tail call double @exp2(double %a1) nounwind readnone
store double %exp1, double* %b, align 8
%idx2 = getelementptr inbounds double, double* %b, i64 1
store double %exp2, double* %idx2, align 8
ret void
}
; No fast-math-flags are required to convert sqrt library calls to an intrinsic.
; We just need to know that errno is not set (readnone).
define void @sqrt_libm_no_errno(double* %a, double* %b) {
; CHECK-LABEL: @sqrt_libm_no_errno(
; CHECK-NEXT: [[TMP1:%.*]] = bitcast double* %a to <2 x double>*
; CHECK-NEXT: [[TMP2:%.*]] = load <2 x double>, <2 x double>* [[TMP1]], align 8
; CHECK-NEXT: [[TMP3:%.*]] = call <2 x double> @llvm.sqrt.v2f64(<2 x double> [[TMP2]])
; CHECK-NEXT: [[TMP4:%.*]] = bitcast double* %b to <2 x double>*
; CHECK-NEXT: store <2 x double> [[TMP3]], <2 x double>* [[TMP4]], align 8
; CHECK-NEXT: ret void
;
%a0 = load double, double* %a, align 8
%idx1 = getelementptr inbounds double, double* %a, i64 1
%a1 = load double, double* %idx1, align 8
%sqrt1 = tail call double @sqrt(double %a0) nounwind readnone
%sqrt2 = tail call double @sqrt(double %a1) nounwind readnone
store double %sqrt1, double* %b, align 8
%idx2 = getelementptr inbounds double, double* %b, i64 1
store double %sqrt2, double* %idx2, align 8
ret void
}
; The sqrt intrinsic does not set errno, but a non-constant sqrt call might, so this can't vectorize.
; The nnan on the call does not matter because there's no guarantee in the C standard that a negative
; input would result in a nan output ("On a domain error, the function returns an
; implementation-defined value.")
define void @sqrt_libm_errno(double* %a, double* %b) {
; CHECK-LABEL: @sqrt_libm_errno(
; CHECK-NEXT: [[A0:%.*]] = load double, double* %a, align 8
; CHECK-NEXT: [[IDX1:%.*]] = getelementptr inbounds double, double* %a, i64 1
; CHECK-NEXT: [[A1:%.*]] = load double, double* [[IDX1]], align 8
; CHECK-NEXT: [[SQRT1:%.*]] = tail call nnan double @sqrt(double [[A0]]) #2
; CHECK-NEXT: [[SQRT2:%.*]] = tail call nnan double @sqrt(double [[A1]]) #2
; CHECK-NEXT: store double [[SQRT1]], double* %b, align 8
; CHECK-NEXT: [[IDX2:%.*]] = getelementptr inbounds double, double* %b, i64 1
; CHECK-NEXT: store double [[SQRT2]], double* [[IDX2]], align 8
; CHECK-NEXT: ret void
;
%a0 = load double, double* %a, align 8
%idx1 = getelementptr inbounds double, double* %a, i64 1
%a1 = load double, double* %idx1, align 8
%sqrt1 = tail call nnan double @sqrt(double %a0) nounwind
%sqrt2 = tail call nnan double @sqrt(double %a1) nounwind
store double %sqrt1, double* %b, align 8
%idx2 = getelementptr inbounds double, double* %b, i64 1
store double %sqrt2, double* %idx2, align 8
ret void
}
; Negative test case
define void @round_custom(i64* %a, i64* %b) {
; CHECK-LABEL: @round_custom(
; CHECK-NEXT: [[A0:%.*]] = load i64, i64* %a, align 8
; CHECK-NEXT: [[IDX1:%.*]] = getelementptr inbounds i64, i64* %a, i64 1
; CHECK-NEXT: [[A1:%.*]] = load i64, i64* [[IDX1]], align 8
; CHECK-NEXT: [[ROUND1:%.*]] = tail call i64 @round(i64 [[A0]]) #3
; CHECK-NEXT: [[ROUND2:%.*]] = tail call i64 @round(i64 [[A1]]) #3
; CHECK-NEXT: store i64 [[ROUND1]], i64* %b, align 8
; CHECK-NEXT: [[IDX2:%.*]] = getelementptr inbounds i64, i64* %b, i64 1
; CHECK-NEXT: store i64 [[ROUND2]], i64* [[IDX2]], align 8
; CHECK-NEXT: ret void
;
%a0 = load i64, i64* %a, align 8
%idx1 = getelementptr inbounds i64, i64* %a, i64 1
%a1 = load i64, i64* %idx1, align 8
%round1 = tail call i64 @round(i64 %a0) nounwind readnone
%round2 = tail call i64 @round(i64 %a1) nounwind readnone
store i64 %round1, i64* %b, align 8
%idx2 = getelementptr inbounds i64, i64* %b, i64 1
store i64 %round2, i64* %idx2, align 8
ret void
}
; CHECK: declare <2 x double> @llvm.sin.v2f64(<2 x double>) [[ATTR0:#[0-9]+]]
; CHECK: declare <2 x double> @llvm.cos.v2f64(<2 x double>) [[ATTR0]]
; CHECK: declare <2 x double> @llvm.pow.v2f64(<2 x double>, <2 x double>) [[ATTR0]]
; CHECK: declare <2 x double> @llvm.exp2.v2f64(<2 x double>) [[ATTR0]]
; CHECK: attributes [[ATTR0]] = { nounwind readnone speculatable }

Some files were not shown because too many files have changed in this diff Show More