1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
|
; RUN: opt -mtriple=amdgcn-amd-amdhsa -mattr=+max-private-element-size-4 -load-store-vectorizer -S -o - %s | FileCheck -check-prefixes=ELT4,ALIGNED,ALL %s
; RUN: opt -mtriple=amdgcn-amd-amdhsa -mattr=+max-private-element-size-8 -load-store-vectorizer -S -o - %s | FileCheck -check-prefixes=ELT8,ALIGNED,ALL %s
; RUN: opt -mtriple=amdgcn-amd-amdhsa -mattr=+max-private-element-size-8,+unaligned-scratch-access -load-store-vectorizer -S -o - %s | FileCheck -check-prefix=ELT8-UNALIGNED -check-prefix=ALL %s
; RUN: opt -mtriple=amdgcn-amd-amdhsa -mattr=+max-private-element-size-16 -load-store-vectorizer -S -o - %s | FileCheck -check-prefixes=ELT16,ALIGNED,ALL %s
; RUN: opt -mtriple=amdgcn-amd-amdhsa -mattr=+max-private-element-size-16,+unaligned-scratch-access -load-store-vectorizer -S -o - %s | FileCheck -check-prefix=ELT16-UNALIGNED -check-prefix=ALL %s
target datalayout = "e-p:32:32-p1:64:64-p2:64:64-p3:32:32-p4:64:64-p5:32:32-p24:64:64-i64:64-v16:16-v24:32-v32:32-v48:64-v96:128-v192:256-v256:256-v512:512-v1024:1024-v2048:2048-n32:64"
; ALL-LABEL: @merge_private_store_4_vector_elts_loads_v4i32
; ALIGNED: store i32
; ALIGNED: store i32
; ALIGNED: store i32
; ALIGNED: store i32
; ELT8-UNALIGNED: store <2 x i32>
; ELT8-UNALIGNED: store <2 x i32>
; ELT16-UNALIGNED: store <4 x i32>
define void @merge_private_store_4_vector_elts_loads_v4i32(i32* %out) #0 {
%out.gep.1 = getelementptr i32, i32* %out, i32 1
%out.gep.2 = getelementptr i32, i32* %out, i32 2
%out.gep.3 = getelementptr i32, i32* %out, i32 3
store i32 9, i32* %out
store i32 1, i32* %out.gep.1
store i32 23, i32* %out.gep.2
store i32 19, i32* %out.gep.3
ret void
}
; ALL-LABEL: @merge_private_store_4_vector_elts_loads_v4i8(
; ALL: store <4 x i8>
define void @merge_private_store_4_vector_elts_loads_v4i8(i8* %out) #0 {
%out.gep.1 = getelementptr i8, i8* %out, i32 1
%out.gep.2 = getelementptr i8, i8* %out, i32 2
%out.gep.3 = getelementptr i8, i8* %out, i32 3
store i8 9, i8* %out, align 4
store i8 1, i8* %out.gep.1
store i8 23, i8* %out.gep.2
store i8 19, i8* %out.gep.3
ret void
}
; ALL-LABEL: @merge_private_store_4_vector_elts_loads_v2i16(
; ALL: store <2 x i16>
define void @merge_private_store_4_vector_elts_loads_v2i16(i16* %out) #0 {
%out.gep.1 = getelementptr i16, i16* %out, i32 1
store i16 9, i16* %out, align 4
store i16 12, i16* %out.gep.1
ret void
}
attributes #0 = { nounwind }
|
