Skip to content

Commit b3d5fba

Browse files
rj-jesusrj-jesus
committed
Update cost of LoopVectorizationCostModel::getInterleaveGroupCost too.
1 parent 92c8faa commit b3d5fba

File tree

4 files changed

+52
-49
lines changed

4 files changed

+52
-49
lines changed

‎llvm/lib/Transforms/Vectorize/LoopVectorize.cpp‎

Lines changed: 3 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -5306,6 +5306,9 @@ LoopVectorizationCostModel::getInterleaveGroupCost(Instruction *I,
53065306
Group->getAlign(), AS, CostKind, Legal->isMaskRequired(I),
53075307
UseMaskForGaps);
53085308

5309+
// Add the address computation cost.
5310+
Cost += TTI.getAddressComputationCost(WideVecTy);
5311+
53095312
if (Group->isReverse()) {
53105313
// TODO: Add support for reversed masked interleaved access.
53115314
assert(!Legal->isMaskRequired(I) &&

‎llvm/test/Transforms/LoopVectorize/AArch64/interleaved_cost.ll‎

Lines changed: 26 additions & 26 deletions
Original file line numberDiff line numberDiff line change
@@ -13,15 +13,15 @@ entry:
1313
br label %for.body
1414

1515
; VF_8-LABEL: Checking a loop in 'i8_factor_2'
16-
; VF_8: Found an estimated cost of 2 for VF 8 For instruction: %tmp2 = load i8, ptr %tmp0, align 1
16+
; VF_8: Found an estimated cost of 3 for VF 8 For instruction: %tmp2 = load i8, ptr %tmp0, align 1
1717
; VF_8-NEXT: Found an estimated cost of 0 for VF 8 For instruction: %tmp3 = load i8, ptr %tmp1, align 1
1818
; VF_8-NEXT: Found an estimated cost of 0 for VF 8 For instruction: store i8 %tmp2, ptr %tmp0, align 1
19-
; VF_8-NEXT: Found an estimated cost of 2 for VF 8 For instruction: store i8 %tmp3, ptr %tmp1, align 1
19+
; VF_8-NEXT: Found an estimated cost of 3 for VF 8 For instruction: store i8 %tmp3, ptr %tmp1, align 1
2020
; VF_16-LABEL: Checking a loop in 'i8_factor_2'
21-
; VF_16: Found an estimated cost of 2 for VF 16 For instruction: %tmp2 = load i8, ptr %tmp0, align 1
21+
; VF_16: Found an estimated cost of 3 for VF 16 For instruction: %tmp2 = load i8, ptr %tmp0, align 1
2222
; VF_16-NEXT: Found an estimated cost of 0 for VF 16 For instruction: %tmp3 = load i8, ptr %tmp1, align 1
2323
; VF_16-NEXT: Found an estimated cost of 0 for VF 16 For instruction: store i8 %tmp2, ptr %tmp0, align 1
24-
; VF_16-NEXT: Found an estimated cost of 2 for VF 16 For instruction: store i8 %tmp3, ptr %tmp1, align 1
24+
; VF_16-NEXT: Found an estimated cost of 3 for VF 16 For instruction: store i8 %tmp3, ptr %tmp1, align 1
2525
for.body:
2626
%i = phi i64 [ 0, %entry ], [ %i.next, %for.body ]
2727
%tmp0 = getelementptr inbounds %i8.2, ptr %data, i64 %i, i32 0
@@ -44,20 +44,20 @@ entry:
4444
br label %for.body
4545

4646
; VF_4-LABEL: Checking a loop in 'i16_factor_2'
47-
; VF_4: Found an estimated cost of 2 for VF 4 For instruction: %tmp2 = load i16, ptr %tmp0, align 2
47+
; VF_4: Found an estimated cost of 3 for VF 4 For instruction: %tmp2 = load i16, ptr %tmp0, align 2
4848
; VF_4-NEXT: Found an estimated cost of 0 for VF 4 For instruction: %tmp3 = load i16, ptr %tmp1, align 2
4949
; VF_4-NEXT: Found an estimated cost of 0 for VF 4 For instruction: store i16 %tmp2, ptr %tmp0, align 2
50-
; VF_4-NEXT: Found an estimated cost of 2 for VF 4 For instruction: store i16 %tmp3, ptr %tmp1, align 2
50+
; VF_4-NEXT: Found an estimated cost of 3 for VF 4 For instruction: store i16 %tmp3, ptr %tmp1, align 2
5151
; VF_8-LABEL: Checking a loop in 'i16_factor_2'
52-
; VF_8: Found an estimated cost of 2 for VF 8 For instruction: %tmp2 = load i16, ptr %tmp0, align 2
52+
; VF_8: Found an estimated cost of 3 for VF 8 For instruction: %tmp2 = load i16, ptr %tmp0, align 2
5353
; VF_8-NEXT: Found an estimated cost of 0 for VF 8 For instruction: %tmp3 = load i16, ptr %tmp1, align 2
5454
; VF_8-NEXT: Found an estimated cost of 0 for VF 8 For instruction: store i16 %tmp2, ptr %tmp0, align 2
55-
; VF_8-NEXT: Found an estimated cost of 2 for VF 8 For instruction: store i16 %tmp3, ptr %tmp1, align 2
55+
; VF_8-NEXT: Found an estimated cost of 3 for VF 8 For instruction: store i16 %tmp3, ptr %tmp1, align 2
5656
; VF_16-LABEL: Checking a loop in 'i16_factor_2'
57-
; VF_16: Found an estimated cost of 4 for VF 16 For instruction: %tmp2 = load i16, ptr %tmp0, align 2
57+
; VF_16: Found an estimated cost of 5 for VF 16 For instruction: %tmp2 = load i16, ptr %tmp0, align 2
5858
; VF_16-NEXT: Found an estimated cost of 0 for VF 16 For instruction: %tmp3 = load i16, ptr %tmp1, align 2
5959
; VF_16-NEXT: Found an estimated cost of 0 for VF 16 For instruction: store i16 %tmp2, ptr %tmp0, align 2
60-
; VF_16-NEXT: Found an estimated cost of 4 for VF 16 For instruction: store i16 %tmp3, ptr %tmp1, align 2
60+
; VF_16-NEXT: Found an estimated cost of 5 for VF 16 For instruction: store i16 %tmp3, ptr %tmp1, align 2
6161
for.body:
6262
%i = phi i64 [ 0, %entry ], [ %i.next, %for.body ]
6363
%tmp0 = getelementptr inbounds %i16.2, ptr %data, i64 %i, i32 0
@@ -80,25 +80,25 @@ entry:
8080
br label %for.body
8181

8282
; VF_2-LABEL: Checking a loop in 'i32_factor_2'
83-
; VF_2: Found an estimated cost of 2 for VF 2 For instruction: %tmp2 = load i32, ptr %tmp0, align 4
83+
; VF_2: Found an estimated cost of 3 for VF 2 For instruction: %tmp2 = load i32, ptr %tmp0, align 4
8484
; VF_2-NEXT: Found an estimated cost of 0 for VF 2 For instruction: %tmp3 = load i32, ptr %tmp1, align 4
8585
; VF_2-NEXT: Found an estimated cost of 0 for VF 2 For instruction: store i32 %tmp2, ptr %tmp0, align 4
86-
; VF_2-NEXT: Found an estimated cost of 2 for VF 2 For instruction: store i32 %tmp3, ptr %tmp1, align 4
86+
; VF_2-NEXT: Found an estimated cost of 3 for VF 2 For instruction: store i32 %tmp3, ptr %tmp1, align 4
8787
; VF_4-LABEL: Checking a loop in 'i32_factor_2'
88-
; VF_4: Found an estimated cost of 2 for VF 4 For instruction: %tmp2 = load i32, ptr %tmp0, align 4
88+
; VF_4: Found an estimated cost of 3 for VF 4 For instruction: %tmp2 = load i32, ptr %tmp0, align 4
8989
; VF_4-NEXT: Found an estimated cost of 0 for VF 4 For instruction: %tmp3 = load i32, ptr %tmp1, align 4
9090
; VF_4-NEXT: Found an estimated cost of 0 for VF 4 For instruction: store i32 %tmp2, ptr %tmp0, align 4
91-
; VF_4-NEXT: Found an estimated cost of 2 for VF 4 For instruction: store i32 %tmp3, ptr %tmp1, align 4
91+
; VF_4-NEXT: Found an estimated cost of 3 for VF 4 For instruction: store i32 %tmp3, ptr %tmp1, align 4
9292
; VF_8-LABEL: Checking a loop in 'i32_factor_2'
93-
; VF_8: Found an estimated cost of 4 for VF 8 For instruction: %tmp2 = load i32, ptr %tmp0, align 4
93+
; VF_8: Found an estimated cost of 5 for VF 8 For instruction: %tmp2 = load i32, ptr %tmp0, align 4
9494
; VF_8-NEXT: Found an estimated cost of 0 for VF 8 For instruction: %tmp3 = load i32, ptr %tmp1, align 4
9595
; VF_8-NEXT: Found an estimated cost of 0 for VF 8 For instruction: store i32 %tmp2, ptr %tmp0, align 4
96-
; VF_8-NEXT: Found an estimated cost of 4 for VF 8 For instruction: store i32 %tmp3, ptr %tmp1, align 4
96+
; VF_8-NEXT: Found an estimated cost of 5 for VF 8 For instruction: store i32 %tmp3, ptr %tmp1, align 4
9797
; VF_16-LABEL: Checking a loop in 'i32_factor_2'
98-
; VF_16: Found an estimated cost of 8 for VF 16 For instruction: %tmp2 = load i32, ptr %tmp0, align 4
98+
; VF_16: Found an estimated cost of 9 for VF 16 For instruction: %tmp2 = load i32, ptr %tmp0, align 4
9999
; VF_16-NEXT: Found an estimated cost of 0 for VF 16 For instruction: %tmp3 = load i32, ptr %tmp1, align 4
100100
; VF_16-NEXT: Found an estimated cost of 0 for VF 16 For instruction: store i32 %tmp2, ptr %tmp0, align 4
101-
; VF_16-NEXT: Found an estimated cost of 8 for VF 16 For instruction: store i32 %tmp3, ptr %tmp1, align 4
101+
; VF_16-NEXT: Found an estimated cost of 9 for VF 16 For instruction: store i32 %tmp3, ptr %tmp1, align 4
102102
for.body:
103103
%i = phi i64 [ 0, %entry ], [ %i.next, %for.body ]
104104
%tmp0 = getelementptr inbounds %i32.2, ptr %data, i64 %i, i32 0
@@ -121,25 +121,25 @@ entry:
121121
br label %for.body
122122

123123
; VF_2-LABEL: Checking a loop in 'i64_factor_2'
124-
; VF_2: Found an estimated cost of 2 for VF 2 For instruction: %tmp2 = load i64, ptr %tmp0, align 8
124+
; VF_2: Found an estimated cost of 3 for VF 2 For instruction: %tmp2 = load i64, ptr %tmp0, align 8
125125
; VF_2-NEXT: Found an estimated cost of 0 for VF 2 For instruction: %tmp3 = load i64, ptr %tmp1, align 8
126126
; VF_2-NEXT: Found an estimated cost of 0 for VF 2 For instruction: store i64 %tmp2, ptr %tmp0, align 8
127-
; VF_2-NEXT: Found an estimated cost of 2 for VF 2 For instruction: store i64 %tmp3, ptr %tmp1, align 8
127+
; VF_2-NEXT: Found an estimated cost of 3 for VF 2 For instruction: store i64 %tmp3, ptr %tmp1, align 8
128128
; VF_4-LABEL: Checking a loop in 'i64_factor_2'
129-
; VF_4: Found an estimated cost of 4 for VF 4 For instruction: %tmp2 = load i64, ptr %tmp0, align 8
129+
; VF_4: Found an estimated cost of 5 for VF 4 For instruction: %tmp2 = load i64, ptr %tmp0, align 8
130130
; VF_4-NEXT: Found an estimated cost of 0 for VF 4 For instruction: %tmp3 = load i64, ptr %tmp1, align 8
131131
; VF_4-NEXT: Found an estimated cost of 0 for VF 4 For instruction: store i64 %tmp2, ptr %tmp0, align 8
132-
; VF_4-NEXT: Found an estimated cost of 4 for VF 4 For instruction: store i64 %tmp3, ptr %tmp1, align 8
132+
; VF_4-NEXT: Found an estimated cost of 5 for VF 4 For instruction: store i64 %tmp3, ptr %tmp1, align 8
133133
; VF_8-LABEL: Checking a loop in 'i64_factor_2'
134-
; VF_8: Found an estimated cost of 8 for VF 8 For instruction: %tmp2 = load i64, ptr %tmp0, align 8
134+
; VF_8: Found an estimated cost of 9 for VF 8 For instruction: %tmp2 = load i64, ptr %tmp0, align 8
135135
; VF_8-NEXT: Found an estimated cost of 0 for VF 8 For instruction: %tmp3 = load i64, ptr %tmp1, align 8
136136
; VF_8-NEXT: Found an estimated cost of 0 for VF 8 For instruction: store i64 %tmp2, ptr %tmp0, align 8
137-
; VF_8-NEXT: Found an estimated cost of 8 for VF 8 For instruction: store i64 %tmp3, ptr %tmp1, align 8
137+
; VF_8-NEXT: Found an estimated cost of 9 for VF 8 For instruction: store i64 %tmp3, ptr %tmp1, align 8
138138
; VF_16-LABEL: Checking a loop in 'i64_factor_2'
139-
; VF_16: Found an estimated cost of 16 for VF 16 For instruction: %tmp2 = load i64, ptr %tmp0, align 8
139+
; VF_16: Found an estimated cost of 17 for VF 16 For instruction: %tmp2 = load i64, ptr %tmp0, align 8
140140
; VF_16-NEXT: Found an estimated cost of 0 for VF 16 For instruction: %tmp3 = load i64, ptr %tmp1, align 8
141141
; VF_16-NEXT: Found an estimated cost of 0 for VF 16 For instruction: store i64 %tmp2, ptr %tmp0, align 8
142-
; VF_16-NEXT: Found an estimated cost of 16 for VF 16 For instruction: store i64 %tmp3, ptr %tmp1, align 8
142+
; VF_16-NEXT: Found an estimated cost of 17 for VF 16 For instruction: store i64 %tmp3, ptr %tmp1, align 8
143143
for.body:
144144
%i = phi i64 [ 0, %entry ], [ %i.next, %for.body ]
145145
%tmp0 = getelementptr inbounds %i64.2, ptr %data, i64 %i, i32 0

‎llvm/test/Transforms/LoopVectorize/AArch64/sve-gather-scatter-cost.ll‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -95,7 +95,7 @@ for.cond.cleanup: ; preds = %for.cond.cleanup.lo
9595
}
9696

9797
; CHECK: LV: Checking a loop in 'gather_nxv4i32_stride2'
98-
; CHECK: LV: Found an estimated cost of 2 for VF vscale x 4 For instruction: %0 = load float, ptr %arrayidx, align 4
98+
; CHECK: LV: Found an estimated cost of 3 for VF vscale x 4 For instruction: %0 = load float, ptr %arrayidx, align 4
9999
define void @gather_nxv4i32_stride2(ptr noalias nocapture readonly %a, ptr noalias nocapture readonly %b, i64 %n) #0 {
100100
entry:
101101
br label %for.body

‎llvm/test/Transforms/LoopVectorize/ARM/interleaved_cost.ll‎

Lines changed: 22 additions & 22 deletions
Original file line numberDiff line numberDiff line change
@@ -13,15 +13,15 @@ entry:
1313
br label %for.body
1414

1515
; VF_8-LABEL: Checking a loop in 'i8_factor_2'
16-
; VF_8: Found an estimated cost of 2 for VF 8 For instruction: %tmp2 = load i8, ptr %tmp0, align 1
16+
; VF_8: Found an estimated cost of 3 for VF 8 For instruction: %tmp2 = load i8, ptr %tmp0, align 1
1717
; VF_8-NEXT: Found an estimated cost of 0 for VF 8 For instruction: %tmp3 = load i8, ptr %tmp1, align 1
1818
; VF_8-NEXT: Found an estimated cost of 0 for VF 8 For instruction: store i8 %tmp2, ptr %tmp0, align 1
19-
; VF_8-NEXT: Found an estimated cost of 2 for VF 8 For instruction: store i8 %tmp3, ptr %tmp1, align 1
19+
; VF_8-NEXT: Found an estimated cost of 3 for VF 8 For instruction: store i8 %tmp3, ptr %tmp1, align 1
2020
; VF_16-LABEL: Checking a loop in 'i8_factor_2'
21-
; VF_16: Found an estimated cost of 2 for VF 16 For instruction: %tmp2 = load i8, ptr %tmp0, align 1
21+
; VF_16: Found an estimated cost of 3 for VF 16 For instruction: %tmp2 = load i8, ptr %tmp0, align 1
2222
; VF_16-NEXT: Found an estimated cost of 0 for VF 16 For instruction: %tmp3 = load i8, ptr %tmp1, align 1
2323
; VF_16-NEXT: Found an estimated cost of 0 for VF 16 For instruction: store i8 %tmp2, ptr %tmp0, align 1
24-
; VF_16-NEXT: Found an estimated cost of 2 for VF 16 For instruction: store i8 %tmp3, ptr %tmp1, align 1
24+
; VF_16-NEXT: Found an estimated cost of 3 for VF 16 For instruction: store i8 %tmp3, ptr %tmp1, align 1
2525
for.body:
2626
%i = phi i64 [ 0, %entry ], [ %i.next, %for.body ]
2727
%tmp0 = getelementptr inbounds %i8.2, ptr %data, i64 %i, i32 0
@@ -44,20 +44,20 @@ entry:
4444
br label %for.body
4545

4646
; VF_4-LABEL: Checking a loop in 'i16_factor_2'
47-
; VF_4: Found an estimated cost of 2 for VF 4 For instruction: %tmp2 = load i16, ptr %tmp0, align 2
47+
; VF_4: Found an estimated cost of 3 for VF 4 For instruction: %tmp2 = load i16, ptr %tmp0, align 2
4848
; VF_4-NEXT: Found an estimated cost of 0 for VF 4 For instruction: %tmp3 = load i16, ptr %tmp1, align 2
4949
; VF_4-NEXT: Found an estimated cost of 0 for VF 4 For instruction: store i16 %tmp2, ptr %tmp0, align 2
50-
; VF_4-NEXT: Found an estimated cost of 2 for VF 4 For instruction: store i16 %tmp3, ptr %tmp1, align 2
50+
; VF_4-NEXT: Found an estimated cost of 3 for VF 4 For instruction: store i16 %tmp3, ptr %tmp1, align 2
5151
; VF_8-LABEL: Checking a loop in 'i16_factor_2'
52-
; VF_8: Found an estimated cost of 2 for VF 8 For instruction: %tmp2 = load i16, ptr %tmp0, align 2
52+
; VF_8: Found an estimated cost of 3 for VF 8 For instruction: %tmp2 = load i16, ptr %tmp0, align 2
5353
; VF_8-NEXT: Found an estimated cost of 0 for VF 8 For instruction: %tmp3 = load i16, ptr %tmp1, align 2
5454
; VF_8-NEXT: Found an estimated cost of 0 for VF 8 For instruction: store i16 %tmp2, ptr %tmp0, align 2
55-
; VF_8-NEXT: Found an estimated cost of 2 for VF 8 For instruction: store i16 %tmp3, ptr %tmp1, align 2
55+
; VF_8-NEXT: Found an estimated cost of 3 for VF 8 For instruction: store i16 %tmp3, ptr %tmp1, align 2
5656
; VF_16-LABEL: Checking a loop in 'i16_factor_2'
57-
; VF_16: Found an estimated cost of 4 for VF 16 For instruction: %tmp2 = load i16, ptr %tmp0, align 2
57+
; VF_16: Found an estimated cost of 5 for VF 16 For instruction: %tmp2 = load i16, ptr %tmp0, align 2
5858
; VF_16-NEXT: Found an estimated cost of 0 for VF 16 For instruction: %tmp3 = load i16, ptr %tmp1, align 2
5959
; VF_16-NEXT: Found an estimated cost of 0 for VF 16 For instruction: store i16 %tmp2, ptr %tmp0, align 2
60-
; VF_16-NEXT: Found an estimated cost of 4 for VF 16 For instruction: store i16 %tmp3, ptr %tmp1, align 2
60+
; VF_16-NEXT: Found an estimated cost of 5 for VF 16 For instruction: store i16 %tmp3, ptr %tmp1, align 2
6161
for.body:
6262
%i = phi i64 [ 0, %entry ], [ %i.next, %for.body ]
6363
%tmp0 = getelementptr inbounds %i16.2, ptr %data, i64 %i, i32 0
@@ -80,25 +80,25 @@ entry:
8080
br label %for.body
8181

8282
; VF_2-LABEL: Checking a loop in 'i32_factor_2'
83-
; VF_2: Found an estimated cost of 2 for VF 2 For instruction: %tmp2 = load i32, ptr %tmp0, align 4
83+
; VF_2: Found an estimated cost of 3 for VF 2 For instruction: %tmp2 = load i32, ptr %tmp0, align 4
8484
; VF_2-NEXT: Found an estimated cost of 0 for VF 2 For instruction: %tmp3 = load i32, ptr %tmp1, align 4
8585
; VF_2-NEXT: Found an estimated cost of 0 for VF 2 For instruction: store i32 %tmp2, ptr %tmp0, align 4
86-
; VF_2-NEXT: Found an estimated cost of 2 for VF 2 For instruction: store i32 %tmp3, ptr %tmp1, align 4
86+
; VF_2-NEXT: Found an estimated cost of 3 for VF 2 For instruction: store i32 %tmp3, ptr %tmp1, align 4
8787
; VF_4-LABEL: Checking a loop in 'i32_factor_2'
88-
; VF_4: Found an estimated cost of 2 for VF 4 For instruction: %tmp2 = load i32, ptr %tmp0, align 4
88+
; VF_4: Found an estimated cost of 3 for VF 4 For instruction: %tmp2 = load i32, ptr %tmp0, align 4
8989
; VF_4-NEXT: Found an estimated cost of 0 for VF 4 For instruction: %tmp3 = load i32, ptr %tmp1, align 4
9090
; VF_4-NEXT: Found an estimated cost of 0 for VF 4 For instruction: store i32 %tmp2, ptr %tmp0, align 4
91-
; VF_4-NEXT: Found an estimated cost of 2 for VF 4 For instruction: store i32 %tmp3, ptr %tmp1, align 4
91+
; VF_4-NEXT: Found an estimated cost of 3 for VF 4 For instruction: store i32 %tmp3, ptr %tmp1, align 4
9292
; VF_8-LABEL: Checking a loop in 'i32_factor_2'
93-
; VF_8: Found an estimated cost of 4 for VF 8 For instruction: %tmp2 = load i32, ptr %tmp0, align 4
93+
; VF_8: Found an estimated cost of 5 for VF 8 For instruction: %tmp2 = load i32, ptr %tmp0, align 4
9494
; VF_8-NEXT: Found an estimated cost of 0 for VF 8 For instruction: %tmp3 = load i32, ptr %tmp1, align 4
9595
; VF_8-NEXT: Found an estimated cost of 0 for VF 8 For instruction: store i32 %tmp2, ptr %tmp0, align 4
96-
; VF_8-NEXT: Found an estimated cost of 4 for VF 8 For instruction: store i32 %tmp3, ptr %tmp1, align 4
96+
; VF_8-NEXT: Found an estimated cost of 5 for VF 8 For instruction: store i32 %tmp3, ptr %tmp1, align 4
9797
; VF_16-LABEL: Checking a loop in 'i32_factor_2'
98-
; VF_16: Found an estimated cost of 8 for VF 16 For instruction: %tmp2 = load i32, ptr %tmp0, align 4
98+
; VF_16: Found an estimated cost of 9 for VF 16 For instruction: %tmp2 = load i32, ptr %tmp0, align 4
9999
; VF_16-NEXT: Found an estimated cost of 0 for VF 16 For instruction: %tmp3 = load i32, ptr %tmp1, align 4
100100
; VF_16-NEXT: Found an estimated cost of 0 for VF 16 For instruction: store i32 %tmp2, ptr %tmp0, align 4
101-
; VF_16-NEXT: Found an estimated cost of 8 for VF 16 For instruction: store i32 %tmp3, ptr %tmp1, align 4
101+
; VF_16-NEXT: Found an estimated cost of 9 for VF 16 For instruction: store i32 %tmp3, ptr %tmp1, align 4
102102
for.body:
103103
%i = phi i64 [ 0, %entry ], [ %i.next, %for.body ]
104104
%tmp0 = getelementptr inbounds %i32.2, ptr %data, i64 %i, i32 0
@@ -121,15 +121,15 @@ entry:
121121
br label %for.body
122122

123123
; VF_4-LABEL: Checking a loop in 'half_factor_2'
124-
; VF_4: Found an estimated cost of 40 for VF 4 For instruction: %tmp2 = load half, ptr %tmp0, align 2
124+
; VF_4: Found an estimated cost of 41 for VF 4 For instruction: %tmp2 = load half, ptr %tmp0, align 2
125125
; VF_4-NEXT: Found an estimated cost of 0 for VF 4 For instruction: %tmp3 = load half, ptr %tmp1, align 2
126126
; VF_4-NEXT: Found an estimated cost of 0 for VF 4 For instruction: store half %tmp2, ptr %tmp0, align 2
127-
; VF_4-NEXT: Found an estimated cost of 40 for VF 4 For instruction: store half %tmp3, ptr %tmp1, align 2
127+
; VF_4-NEXT: Found an estimated cost of 41 for VF 4 For instruction: store half %tmp3, ptr %tmp1, align 2
128128
; VF_8-LABEL: Checking a loop in 'half_factor_2'
129-
; VF_8: Found an estimated cost of 80 for VF 8 For instruction: %tmp2 = load half, ptr %tmp0, align 2
129+
; VF_8: Found an estimated cost of 81 for VF 8 For instruction: %tmp2 = load half, ptr %tmp0, align 2
130130
; VF_8-NEXT: Found an estimated cost of 0 for VF 8 For instruction: %tmp3 = load half, ptr %tmp1, align 2
131131
; VF_8-NEXT: Found an estimated cost of 0 for VF 8 For instruction: store half %tmp2, ptr %tmp0, align 2
132-
; VF_8-NEXT: Found an estimated cost of 80 for VF 8 For instruction: store half %tmp3, ptr %tmp1, align 2
132+
; VF_8-NEXT: Found an estimated cost of 81 for VF 8 For instruction: store half %tmp3, ptr %tmp1, align 2
133133
for.body:
134134
%i = phi i64 [ 0, %entry ], [ %i.next, %for.body ]
135135
%tmp0 = getelementptr inbounds %half.2, ptr %data, i64 %i, i32 0

0 commit comments

Comments
 (0)