aboutsummaryrefslogtreecommitdiffstats
path: root/test/Transforms/LoopVectorize/ARM
diff options
context:
space:
mode:
Diffstat (limited to 'test/Transforms/LoopVectorize/ARM')
-rw-r--r--test/Transforms/LoopVectorize/ARM/arm-unroll.ll8
-rw-r--r--test/Transforms/LoopVectorize/ARM/gather-cost.ll28
-rw-r--r--test/Transforms/LoopVectorize/ARM/gcc-examples.ll16
-rw-r--r--test/Transforms/LoopVectorize/ARM/mul-cast-vect.ll24
-rw-r--r--test/Transforms/LoopVectorize/ARM/width-detect.ll8
5 files changed, 42 insertions, 42 deletions
diff --git a/test/Transforms/LoopVectorize/ARM/arm-unroll.ll b/test/Transforms/LoopVectorize/ARM/arm-unroll.ll
index 8843fc2..7b09913 100644
--- a/test/Transforms/LoopVectorize/ARM/arm-unroll.ll
+++ b/test/Transforms/LoopVectorize/ARM/arm-unroll.ll
@@ -20,8 +20,8 @@ define i32 @foo(i32* nocapture %A, i32 %n) nounwind readonly ssp {
.lr.ph: ; preds = %0, %.lr.ph
%i.02 = phi i32 [ %5, %.lr.ph ], [ 0, %0 ]
%sum.01 = phi i32 [ %4, %.lr.ph ], [ 0, %0 ]
- %2 = getelementptr inbounds i32* %A, i32 %i.02
- %3 = load i32* %2, align 4
+ %2 = getelementptr inbounds i32, i32* %A, i32 %i.02
+ %3 = load i32, i32* %2, align 4
%4 = add nsw i32 %3, %sum.01
%5 = add nsw i32 %i.02, 1
%exitcond = icmp eq i32 %5, %n
@@ -48,8 +48,8 @@ define i32 @register_limit(i32* nocapture %A, i32 %n) {
%sum.04 = phi i32 [ %8, %.lr.ph ], [ 0, %0 ]
%sum.05 = phi i32 [ %9, %.lr.ph ], [ 0, %0 ]
%sum.06 = phi i32 [ %10, %.lr.ph ], [ 0, %0 ]
- %2 = getelementptr inbounds i32* %A, i32 %i.02
- %3 = load i32* %2, align 4
+ %2 = getelementptr inbounds i32, i32* %A, i32 %i.02
+ %3 = load i32, i32* %2, align 4
%4 = add nsw i32 %3, %sum.01
%5 = add nsw i32 %i.02, 1
%6 = add nsw i32 %3, %sum.02
diff --git a/test/Transforms/LoopVectorize/ARM/gather-cost.ll b/test/Transforms/LoopVectorize/ARM/gather-cost.ll
index 239a28f..f14a8cc 100644
--- a/test/Transforms/LoopVectorize/ARM/gather-cost.ll
+++ b/test/Transforms/LoopVectorize/ARM/gather-cost.ll
@@ -33,32 +33,32 @@ for.body:
%b.054 = phi float [ 0.000000e+00, %for.body.lr.ph ], [ %add30, %for.body ]
%add = add i32 %v.055, %offset
%mul = mul i32 %add, 3
- %arrayidx = getelementptr inbounds [1536 x float]* @src_data, i32 0, i32 %mul
- %0 = load float* %arrayidx, align 4
- %arrayidx2 = getelementptr inbounds [512 x float]* @kernel, i32 0, i32 %v.055
- %1 = load float* %arrayidx2, align 4
+ %arrayidx = getelementptr inbounds [1536 x float], [1536 x float]* @src_data, i32 0, i32 %mul
+ %0 = load float, float* %arrayidx, align 4
+ %arrayidx2 = getelementptr inbounds [512 x float], [512 x float]* @kernel, i32 0, i32 %v.055
+ %1 = load float, float* %arrayidx2, align 4
%mul3 = fmul fast float %0, %1
- %arrayidx4 = getelementptr inbounds [512 x float]* @kernel2, i32 0, i32 %v.055
- %2 = load float* %arrayidx4, align 4
+ %arrayidx4 = getelementptr inbounds [512 x float], [512 x float]* @kernel2, i32 0, i32 %v.055
+ %2 = load float, float* %arrayidx4, align 4
%mul5 = fmul fast float %mul3, %2
- %arrayidx6 = getelementptr inbounds [512 x float]* @kernel3, i32 0, i32 %v.055
- %3 = load float* %arrayidx6, align 4
+ %arrayidx6 = getelementptr inbounds [512 x float], [512 x float]* @kernel3, i32 0, i32 %v.055
+ %3 = load float, float* %arrayidx6, align 4
%mul7 = fmul fast float %mul5, %3
- %arrayidx8 = getelementptr inbounds [512 x float]* @kernel4, i32 0, i32 %v.055
- %4 = load float* %arrayidx8, align 4
+ %arrayidx8 = getelementptr inbounds [512 x float], [512 x float]* @kernel4, i32 0, i32 %v.055
+ %4 = load float, float* %arrayidx8, align 4
%mul9 = fmul fast float %mul7, %4
%add10 = fadd fast float %r.057, %mul9
%arrayidx.sum = add i32 %mul, 1
- %arrayidx11 = getelementptr inbounds [1536 x float]* @src_data, i32 0, i32 %arrayidx.sum
- %5 = load float* %arrayidx11, align 4
+ %arrayidx11 = getelementptr inbounds [1536 x float], [1536 x float]* @src_data, i32 0, i32 %arrayidx.sum
+ %5 = load float, float* %arrayidx11, align 4
%mul13 = fmul fast float %1, %5
%mul15 = fmul fast float %2, %mul13
%mul17 = fmul fast float %3, %mul15
%mul19 = fmul fast float %4, %mul17
%add20 = fadd fast float %g.056, %mul19
%arrayidx.sum52 = add i32 %mul, 2
- %arrayidx21 = getelementptr inbounds [1536 x float]* @src_data, i32 0, i32 %arrayidx.sum52
- %6 = load float* %arrayidx21, align 4
+ %arrayidx21 = getelementptr inbounds [1536 x float], [1536 x float]* @src_data, i32 0, i32 %arrayidx.sum52
+ %6 = load float, float* %arrayidx21, align 4
%mul23 = fmul fast float %1, %6
%mul25 = fmul fast float %2, %mul23
%mul27 = fmul fast float %3, %mul25
diff --git a/test/Transforms/LoopVectorize/ARM/gcc-examples.ll b/test/Transforms/LoopVectorize/ARM/gcc-examples.ll
index f2bd0ac..783156d 100644
--- a/test/Transforms/LoopVectorize/ARM/gcc-examples.ll
+++ b/test/Transforms/LoopVectorize/ARM/gcc-examples.ll
@@ -18,12 +18,12 @@ define void @example1() nounwind uwtable ssp {
; <label>:1 ; preds = %1, %0
%indvars.iv = phi i64 [ 0, %0 ], [ %indvars.iv.next, %1 ]
- %2 = getelementptr inbounds [2048 x i32]* @b, i64 0, i64 %indvars.iv
- %3 = load i32* %2, align 4
- %4 = getelementptr inbounds [2048 x i32]* @c, i64 0, i64 %indvars.iv
- %5 = load i32* %4, align 4
+ %2 = getelementptr inbounds [2048 x i32], [2048 x i32]* @b, i64 0, i64 %indvars.iv
+ %3 = load i32, i32* %2, align 4
+ %4 = getelementptr inbounds [2048 x i32], [2048 x i32]* @c, i64 0, i64 %indvars.iv
+ %5 = load i32, i32* %4, align 4
%6 = add nsw i32 %5, %3
- %7 = getelementptr inbounds [2048 x i32]* @a, i64 0, i64 %indvars.iv
+ %7 = getelementptr inbounds [2048 x i32], [2048 x i32]* @a, i64 0, i64 %indvars.iv
store i32 %6, i32* %7, align 4
%indvars.iv.next = add i64 %indvars.iv, 1
%lftr.wideiv = trunc i64 %indvars.iv.next to i32
@@ -44,10 +44,10 @@ define void @example10b(i16* noalias nocapture %sa, i16* noalias nocapture %sb,
; <label>:1 ; preds = %1, %0
%indvars.iv = phi i64 [ 0, %0 ], [ %indvars.iv.next, %1 ]
- %2 = getelementptr inbounds i16* %sb, i64 %indvars.iv
- %3 = load i16* %2, align 2
+ %2 = getelementptr inbounds i16, i16* %sb, i64 %indvars.iv
+ %3 = load i16, i16* %2, align 2
%4 = sext i16 %3 to i32
- %5 = getelementptr inbounds i32* %ia, i64 %indvars.iv
+ %5 = getelementptr inbounds i32, i32* %ia, i64 %indvars.iv
store i32 %4, i32* %5, align 4
%indvars.iv.next = add i64 %indvars.iv, 1
%lftr.wideiv = trunc i64 %indvars.iv.next to i32
diff --git a/test/Transforms/LoopVectorize/ARM/mul-cast-vect.ll b/test/Transforms/LoopVectorize/ARM/mul-cast-vect.ll
index d2e3de2..e88fcca 100644
--- a/test/Transforms/LoopVectorize/ARM/mul-cast-vect.ll
+++ b/test/Transforms/LoopVectorize/ARM/mul-cast-vect.ll
@@ -16,9 +16,9 @@ target triple = "armv7--linux-gnueabihf"
define void @direct(%T432* %loadaddr, %T432* %loadaddr2, %T432* %storeaddr) {
; COST: function 'direct':
- %v0 = load %T432* %loadaddr
+ %v0 = load %T432, %T432* %loadaddr
; ASM: vld1.64
- %v1 = load %T432* %loadaddr2
+ %v1 = load %T432, %T432* %loadaddr2
; ASM: vld1.64
%r3 = mul %T432 %v0, %v1
; COST: cost of 2 for instruction: {{.*}} mul <4 x i32>
@@ -30,9 +30,9 @@ define void @direct(%T432* %loadaddr, %T432* %loadaddr2, %T432* %storeaddr) {
define void @ups1632(%T416* %loadaddr, %T416* %loadaddr2, %T432* %storeaddr) {
; COST: function 'ups1632':
- %v0 = load %T416* %loadaddr
+ %v0 = load %T416, %T416* %loadaddr
; ASM: vldr
- %v1 = load %T416* %loadaddr2
+ %v1 = load %T416, %T416* %loadaddr2
; ASM: vldr
%r1 = sext %T416 %v0 to %T432
%r2 = sext %T416 %v1 to %T432
@@ -47,9 +47,9 @@ define void @ups1632(%T416* %loadaddr, %T416* %loadaddr2, %T432* %storeaddr) {
define void @upu1632(%T416* %loadaddr, %T416* %loadaddr2, %T432* %storeaddr) {
; COST: function 'upu1632':
- %v0 = load %T416* %loadaddr
+ %v0 = load %T416, %T416* %loadaddr
; ASM: vldr
- %v1 = load %T416* %loadaddr2
+ %v1 = load %T416, %T416* %loadaddr2
; ASM: vldr
%r1 = zext %T416 %v0 to %T432
%r2 = zext %T416 %v1 to %T432
@@ -64,9 +64,9 @@ define void @upu1632(%T416* %loadaddr, %T416* %loadaddr2, %T432* %storeaddr) {
define void @ups3264(%T232* %loadaddr, %T232* %loadaddr2, %T264* %storeaddr) {
; COST: function 'ups3264':
- %v0 = load %T232* %loadaddr
+ %v0 = load %T232, %T232* %loadaddr
; ASM: vldr
- %v1 = load %T232* %loadaddr2
+ %v1 = load %T232, %T232* %loadaddr2
; ASM: vldr
%r3 = mul %T232 %v0, %v1
; ASM: vmul.i32
@@ -81,9 +81,9 @@ define void @ups3264(%T232* %loadaddr, %T232* %loadaddr2, %T264* %storeaddr) {
define void @upu3264(%T232* %loadaddr, %T232* %loadaddr2, %T264* %storeaddr) {
; COST: function 'upu3264':
- %v0 = load %T232* %loadaddr
+ %v0 = load %T232, %T232* %loadaddr
; ASM: vldr
- %v1 = load %T232* %loadaddr2
+ %v1 = load %T232, %T232* %loadaddr2
; ASM: vldr
%r3 = mul %T232 %v0, %v1
; ASM: vmul.i32
@@ -98,9 +98,9 @@ define void @upu3264(%T232* %loadaddr, %T232* %loadaddr2, %T264* %storeaddr) {
define void @dn3216(%T432* %loadaddr, %T432* %loadaddr2, %T416* %storeaddr) {
; COST: function 'dn3216':
- %v0 = load %T432* %loadaddr
+ %v0 = load %T432, %T432* %loadaddr
; ASM: vld1.64
- %v1 = load %T432* %loadaddr2
+ %v1 = load %T432, %T432* %loadaddr2
; ASM: vld1.64
%r3 = mul %T432 %v0, %v1
; ASM: vmul.i32
diff --git a/test/Transforms/LoopVectorize/ARM/width-detect.ll b/test/Transforms/LoopVectorize/ARM/width-detect.ll
index 99d7fa7..66d2556 100644
--- a/test/Transforms/LoopVectorize/ARM/width-detect.ll
+++ b/test/Transforms/LoopVectorize/ARM/width-detect.ll
@@ -13,8 +13,8 @@ define float @foo_F32(float* nocapture %A, i32 %n) nounwind uwtable readonly ssp
.lr.ph: ; preds = %0, %.lr.ph
%indvars.iv = phi i64 [ %indvars.iv.next, %.lr.ph ], [ 0, %0 ]
%prod.01 = phi float [ %4, %.lr.ph ], [ 0.000000e+00, %0 ]
- %2 = getelementptr inbounds float* %A, i64 %indvars.iv
- %3 = load float* %2, align 8
+ %2 = getelementptr inbounds float, float* %A, i64 %indvars.iv
+ %3 = load float, float* %2, align 8
%4 = fmul fast float %prod.01, %3
%indvars.iv.next = add i64 %indvars.iv, 1
%lftr.wideiv = trunc i64 %indvars.iv.next to i32
@@ -36,8 +36,8 @@ define signext i8 @foo_I8(i8* nocapture %A, i32 %n) nounwind uwtable readonly ss
.lr.ph: ; preds = %0, %.lr.ph
%indvars.iv = phi i64 [ %indvars.iv.next, %.lr.ph ], [ 0, %0 ]
%red.01 = phi i8 [ %4, %.lr.ph ], [ 0, %0 ]
- %2 = getelementptr inbounds i8* %A, i64 %indvars.iv
- %3 = load i8* %2, align 1
+ %2 = getelementptr inbounds i8, i8* %A, i64 %indvars.iv
+ %3 = load i8, i8* %2, align 1
%4 = xor i8 %3, %red.01
%indvars.iv.next = add i64 %indvars.iv, 1
%lftr.wideiv = trunc i64 %indvars.iv.next to i32