Skip to content

Commit 3df9dbf

Browse files
Mel-ChenMel-Chen
committed
Rebase
1 parent ac56a99 commit 3df9dbf

File tree

3 files changed

+59
-59
lines changed

3 files changed

+59
-59
lines changed

‎llvm/test/Transforms/LoopVectorize/iv-select-cmp-trunc.ll

Lines changed: 21 additions & 21 deletions
Original file line numberDiff line numberDiff line change
@@ -95,12 +95,12 @@ define i32 @select_icmp_const_truncated_iv_widened_exit(ptr %a, i32 %n) {
9595
; CHECK-VF4IC4-NEXT: [[TMP6:%.*]] = getelementptr inbounds i64, ptr [[A]], i64 [[TMP2]]
9696
; CHECK-VF4IC4-NEXT: [[TMP7:%.*]] = getelementptr inbounds i64, ptr [[A]], i64 [[TMP3]]
9797
; CHECK-VF4IC4-NEXT: [[TMP8:%.*]] = getelementptr inbounds i64, ptr [[TMP4]], i32 0
98-
; CHECK-VF4IC4-NEXT: [[WIDE_LOAD:%.*]] = load <4 x i64>, ptr [[TMP8]], align 8
9998
; CHECK-VF4IC4-NEXT: [[TMP9:%.*]] = getelementptr inbounds i64, ptr [[TMP4]], i32 4
100-
; CHECK-VF4IC4-NEXT: [[WIDE_LOAD7:%.*]] = load <4 x i64>, ptr [[TMP9]], align 8
10199
; CHECK-VF4IC4-NEXT: [[TMP10:%.*]] = getelementptr inbounds i64, ptr [[TMP4]], i32 8
102-
; CHECK-VF4IC4-NEXT: [[WIDE_LOAD8:%.*]] = load <4 x i64>, ptr [[TMP10]], align 8
103100
; CHECK-VF4IC4-NEXT: [[TMP11:%.*]] = getelementptr inbounds i64, ptr [[TMP4]], i32 12
101+
; CHECK-VF4IC4-NEXT: [[WIDE_LOAD:%.*]] = load <4 x i64>, ptr [[TMP8]], align 8
102+
; CHECK-VF4IC4-NEXT: [[WIDE_LOAD7:%.*]] = load <4 x i64>, ptr [[TMP9]], align 8
103+
; CHECK-VF4IC4-NEXT: [[WIDE_LOAD8:%.*]] = load <4 x i64>, ptr [[TMP10]], align 8
104104
; CHECK-VF4IC4-NEXT: [[WIDE_LOAD9:%.*]] = load <4 x i64>, ptr [[TMP11]], align 8
105105
; CHECK-VF4IC4-NEXT: [[TMP12:%.*]] = icmp sgt <4 x i64> [[WIDE_LOAD]], <i64 3, i64 3, i64 3, i64 3>
106106
; CHECK-VF4IC4-NEXT: [[TMP13:%.*]] = icmp sgt <4 x i64> [[WIDE_LOAD7]], <i64 3, i64 3, i64 3, i64 3>
@@ -164,15 +164,15 @@ define i32 @select_icmp_const_truncated_iv_widened_exit(ptr %a, i32 %n) {
164164
; CHECK-VF1IC4-NEXT: [[VEC_PHI1:%.*]] = phi i32 [ -2147483648, [[VECTOR_PH]] ], [ [[TMP22:%.*]], [[VECTOR_BODY]] ]
165165
; CHECK-VF1IC4-NEXT: [[VEC_PHI2:%.*]] = phi i32 [ -2147483648, [[VECTOR_PH]] ], [ [[TMP23:%.*]], [[VECTOR_BODY]] ]
166166
; CHECK-VF1IC4-NEXT: [[VEC_PHI3:%.*]] = phi i32 [ -2147483648, [[VECTOR_PH]] ], [ [[TMP24:%.*]], [[VECTOR_BODY]] ]
167+
; CHECK-VF1IC4-NEXT: [[TMP5:%.*]] = add i64 [[OFFSET_IDX]], 0
168+
; CHECK-VF1IC4-NEXT: [[TMP6:%.*]] = add i64 [[OFFSET_IDX]], 1
169+
; CHECK-VF1IC4-NEXT: [[TMP7:%.*]] = add i64 [[OFFSET_IDX]], 2
170+
; CHECK-VF1IC4-NEXT: [[TMP8:%.*]] = add i64 [[OFFSET_IDX]], 3
167171
; CHECK-VF1IC4-NEXT: [[TMP0:%.*]] = trunc i64 [[OFFSET_IDX]] to i32
168172
; CHECK-VF1IC4-NEXT: [[TMP1:%.*]] = add i32 [[TMP0]], 0
169173
; CHECK-VF1IC4-NEXT: [[TMP2:%.*]] = add i32 [[TMP0]], 1
170174
; CHECK-VF1IC4-NEXT: [[TMP3:%.*]] = add i32 [[TMP0]], 2
171175
; CHECK-VF1IC4-NEXT: [[TMP4:%.*]] = add i32 [[TMP0]], 3
172-
; CHECK-VF1IC4-NEXT: [[TMP5:%.*]] = add i64 [[OFFSET_IDX]], 0
173-
; CHECK-VF1IC4-NEXT: [[TMP6:%.*]] = add i64 [[OFFSET_IDX]], 1
174-
; CHECK-VF1IC4-NEXT: [[TMP7:%.*]] = add i64 [[OFFSET_IDX]], 2
175-
; CHECK-VF1IC4-NEXT: [[TMP8:%.*]] = add i64 [[OFFSET_IDX]], 3
176176
; CHECK-VF1IC4-NEXT: [[TMP9:%.*]] = getelementptr inbounds i64, ptr [[A]], i64 [[TMP5]]
177177
; CHECK-VF1IC4-NEXT: [[TMP10:%.*]] = getelementptr inbounds i64, ptr [[A]], i64 [[TMP6]]
178178
; CHECK-VF1IC4-NEXT: [[TMP11:%.*]] = getelementptr inbounds i64, ptr [[A]], i64 [[TMP7]]
@@ -319,12 +319,12 @@ define i32 @select_icmp_const_truncated_iv_const_exit(ptr %a) {
319319
; CHECK-VF4IC4-NEXT: [[TMP6:%.*]] = getelementptr inbounds i64, ptr [[A]], i64 [[TMP2]]
320320
; CHECK-VF4IC4-NEXT: [[TMP7:%.*]] = getelementptr inbounds i64, ptr [[A]], i64 [[TMP3]]
321321
; CHECK-VF4IC4-NEXT: [[TMP8:%.*]] = getelementptr inbounds i64, ptr [[TMP4]], i32 0
322-
; CHECK-VF4IC4-NEXT: [[WIDE_LOAD:%.*]] = load <4 x i64>, ptr [[TMP8]], align 8
323322
; CHECK-VF4IC4-NEXT: [[TMP9:%.*]] = getelementptr inbounds i64, ptr [[TMP4]], i32 4
324-
; CHECK-VF4IC4-NEXT: [[WIDE_LOAD7:%.*]] = load <4 x i64>, ptr [[TMP9]], align 8
325323
; CHECK-VF4IC4-NEXT: [[TMP10:%.*]] = getelementptr inbounds i64, ptr [[TMP4]], i32 8
326-
; CHECK-VF4IC4-NEXT: [[WIDE_LOAD8:%.*]] = load <4 x i64>, ptr [[TMP10]], align 8
327324
; CHECK-VF4IC4-NEXT: [[TMP11:%.*]] = getelementptr inbounds i64, ptr [[TMP4]], i32 12
325+
; CHECK-VF4IC4-NEXT: [[WIDE_LOAD:%.*]] = load <4 x i64>, ptr [[TMP8]], align 8
326+
; CHECK-VF4IC4-NEXT: [[WIDE_LOAD7:%.*]] = load <4 x i64>, ptr [[TMP9]], align 8
327+
; CHECK-VF4IC4-NEXT: [[WIDE_LOAD8:%.*]] = load <4 x i64>, ptr [[TMP10]], align 8
328328
; CHECK-VF4IC4-NEXT: [[WIDE_LOAD9:%.*]] = load <4 x i64>, ptr [[TMP11]], align 8
329329
; CHECK-VF4IC4-NEXT: [[TMP12:%.*]] = icmp sgt <4 x i64> [[WIDE_LOAD]], <i64 3, i64 3, i64 3, i64 3>
330330
; CHECK-VF4IC4-NEXT: [[TMP13:%.*]] = icmp sgt <4 x i64> [[WIDE_LOAD7]], <i64 3, i64 3, i64 3, i64 3>
@@ -377,15 +377,15 @@ define i32 @select_icmp_const_truncated_iv_const_exit(ptr %a) {
377377
; CHECK-VF1IC4-NEXT: [[VEC_PHI1:%.*]] = phi i32 [ -2147483648, [[VECTOR_PH]] ], [ [[TMP22:%.*]], [[VECTOR_BODY]] ]
378378
; CHECK-VF1IC4-NEXT: [[VEC_PHI2:%.*]] = phi i32 [ -2147483648, [[VECTOR_PH]] ], [ [[TMP23:%.*]], [[VECTOR_BODY]] ]
379379
; CHECK-VF1IC4-NEXT: [[VEC_PHI3:%.*]] = phi i32 [ -2147483648, [[VECTOR_PH]] ], [ [[TMP24:%.*]], [[VECTOR_BODY]] ]
380+
; CHECK-VF1IC4-NEXT: [[TMP5:%.*]] = add i64 [[OFFSET_IDX]], 0
381+
; CHECK-VF1IC4-NEXT: [[TMP6:%.*]] = add i64 [[OFFSET_IDX]], 1
382+
; CHECK-VF1IC4-NEXT: [[TMP7:%.*]] = add i64 [[OFFSET_IDX]], 2
383+
; CHECK-VF1IC4-NEXT: [[TMP8:%.*]] = add i64 [[OFFSET_IDX]], 3
380384
; CHECK-VF1IC4-NEXT: [[TMP0:%.*]] = trunc i64 [[OFFSET_IDX]] to i32
381385
; CHECK-VF1IC4-NEXT: [[TMP1:%.*]] = add i32 [[TMP0]], 0
382386
; CHECK-VF1IC4-NEXT: [[TMP2:%.*]] = add i32 [[TMP0]], 1
383387
; CHECK-VF1IC4-NEXT: [[TMP3:%.*]] = add i32 [[TMP0]], 2
384388
; CHECK-VF1IC4-NEXT: [[TMP4:%.*]] = add i32 [[TMP0]], 3
385-
; CHECK-VF1IC4-NEXT: [[TMP5:%.*]] = add i64 [[OFFSET_IDX]], 0
386-
; CHECK-VF1IC4-NEXT: [[TMP6:%.*]] = add i64 [[OFFSET_IDX]], 1
387-
; CHECK-VF1IC4-NEXT: [[TMP7:%.*]] = add i64 [[OFFSET_IDX]], 2
388-
; CHECK-VF1IC4-NEXT: [[TMP8:%.*]] = add i64 [[OFFSET_IDX]], 3
389389
; CHECK-VF1IC4-NEXT: [[TMP9:%.*]] = getelementptr inbounds i64, ptr [[A]], i64 [[TMP5]]
390390
; CHECK-VF1IC4-NEXT: [[TMP10:%.*]] = getelementptr inbounds i64, ptr [[A]], i64 [[TMP6]]
391391
; CHECK-VF1IC4-NEXT: [[TMP11:%.*]] = getelementptr inbounds i64, ptr [[A]], i64 [[TMP7]]
@@ -522,12 +522,12 @@ define i32 @select_fcmp_max_valid_const_ub(ptr %a) {
522522
; CHECK-VF4IC4-NEXT: [[TMP6:%.*]] = getelementptr inbounds float, ptr [[A]], i64 [[TMP2]]
523523
; CHECK-VF4IC4-NEXT: [[TMP7:%.*]] = getelementptr inbounds float, ptr [[A]], i64 [[TMP3]]
524524
; CHECK-VF4IC4-NEXT: [[TMP8:%.*]] = getelementptr inbounds float, ptr [[TMP4]], i32 0
525-
; CHECK-VF4IC4-NEXT: [[WIDE_LOAD:%.*]] = load <4 x float>, ptr [[TMP8]], align 4
526525
; CHECK-VF4IC4-NEXT: [[TMP9:%.*]] = getelementptr inbounds float, ptr [[TMP4]], i32 4
527-
; CHECK-VF4IC4-NEXT: [[WIDE_LOAD7:%.*]] = load <4 x float>, ptr [[TMP9]], align 4
528526
; CHECK-VF4IC4-NEXT: [[TMP10:%.*]] = getelementptr inbounds float, ptr [[TMP4]], i32 8
529-
; CHECK-VF4IC4-NEXT: [[WIDE_LOAD8:%.*]] = load <4 x float>, ptr [[TMP10]], align 4
530527
; CHECK-VF4IC4-NEXT: [[TMP11:%.*]] = getelementptr inbounds float, ptr [[TMP4]], i32 12
528+
; CHECK-VF4IC4-NEXT: [[WIDE_LOAD:%.*]] = load <4 x float>, ptr [[TMP8]], align 4
529+
; CHECK-VF4IC4-NEXT: [[WIDE_LOAD7:%.*]] = load <4 x float>, ptr [[TMP9]], align 4
530+
; CHECK-VF4IC4-NEXT: [[WIDE_LOAD8:%.*]] = load <4 x float>, ptr [[TMP10]], align 4
531531
; CHECK-VF4IC4-NEXT: [[WIDE_LOAD9:%.*]] = load <4 x float>, ptr [[TMP11]], align 4
532532
; CHECK-VF4IC4-NEXT: [[TMP12:%.*]] = fcmp fast olt <4 x float> [[WIDE_LOAD]], zeroinitializer
533533
; CHECK-VF4IC4-NEXT: [[TMP13:%.*]] = fcmp fast olt <4 x float> [[WIDE_LOAD7]], zeroinitializer
@@ -580,15 +580,15 @@ define i32 @select_fcmp_max_valid_const_ub(ptr %a) {
580580
; CHECK-VF1IC4-NEXT: [[VEC_PHI1:%.*]] = phi i32 [ -2147483648, [[VECTOR_PH]] ], [ [[TMP22:%.*]], [[VECTOR_BODY]] ]
581581
; CHECK-VF1IC4-NEXT: [[VEC_PHI2:%.*]] = phi i32 [ -2147483648, [[VECTOR_PH]] ], [ [[TMP23:%.*]], [[VECTOR_BODY]] ]
582582
; CHECK-VF1IC4-NEXT: [[VEC_PHI3:%.*]] = phi i32 [ -2147483648, [[VECTOR_PH]] ], [ [[TMP24:%.*]], [[VECTOR_BODY]] ]
583+
; CHECK-VF1IC4-NEXT: [[TMP5:%.*]] = add i64 [[OFFSET_IDX]], 0
584+
; CHECK-VF1IC4-NEXT: [[TMP6:%.*]] = add i64 [[OFFSET_IDX]], 1
585+
; CHECK-VF1IC4-NEXT: [[TMP7:%.*]] = add i64 [[OFFSET_IDX]], 2
586+
; CHECK-VF1IC4-NEXT: [[TMP8:%.*]] = add i64 [[OFFSET_IDX]], 3
583587
; CHECK-VF1IC4-NEXT: [[TMP0:%.*]] = trunc i64 [[OFFSET_IDX]] to i32
584588
; CHECK-VF1IC4-NEXT: [[TMP1:%.*]] = add i32 [[TMP0]], 0
585589
; CHECK-VF1IC4-NEXT: [[TMP2:%.*]] = add i32 [[TMP0]], 1
586590
; CHECK-VF1IC4-NEXT: [[TMP3:%.*]] = add i32 [[TMP0]], 2
587591
; CHECK-VF1IC4-NEXT: [[TMP4:%.*]] = add i32 [[TMP0]], 3
588-
; CHECK-VF1IC4-NEXT: [[TMP5:%.*]] = add i64 [[OFFSET_IDX]], 0
589-
; CHECK-VF1IC4-NEXT: [[TMP6:%.*]] = add i64 [[OFFSET_IDX]], 1
590-
; CHECK-VF1IC4-NEXT: [[TMP7:%.*]] = add i64 [[OFFSET_IDX]], 2
591-
; CHECK-VF1IC4-NEXT: [[TMP8:%.*]] = add i64 [[OFFSET_IDX]], 3
592592
; CHECK-VF1IC4-NEXT: [[TMP9:%.*]] = getelementptr inbounds float, ptr [[A]], i64 [[TMP5]]
593593
; CHECK-VF1IC4-NEXT: [[TMP10:%.*]] = getelementptr inbounds float, ptr [[A]], i64 [[TMP6]]
594594
; CHECK-VF1IC4-NEXT: [[TMP11:%.*]] = getelementptr inbounds float, ptr [[A]], i64 [[TMP7]]

0 commit comments

Comments
 (0)