arm64-sli-sri-opt.ll
2.23 KB
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
; RUN: llc < %s -aarch64-shift-insert-generation=true -mtriple=arm64-eabi -aarch64-neon-syntax=apple | FileCheck %s
define void @testLeftGood(<16 x i8> %src1, <16 x i8> %src2, <16 x i8>* %dest) nounwind {
; CHECK-LABEL: testLeftGood:
; CHECK: sli.16b v0, v1, #3
%and.i = and <16 x i8> %src1, <i8 252, i8 252, i8 252, i8 252, i8 252, i8 252, i8 252, i8 252, i8 252, i8 252, i8 252, i8 252, i8 252, i8 252, i8 252, i8 252>
%vshl_n = shl <16 x i8> %src2, <i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3>
%result = or <16 x i8> %and.i, %vshl_n
store <16 x i8> %result, <16 x i8>* %dest, align 16
ret void
}
define void @testLeftBad(<16 x i8> %src1, <16 x i8> %src2, <16 x i8>* %dest) nounwind {
; CHECK-LABEL: testLeftBad:
; CHECK-NOT: sli
%and.i = and <16 x i8> %src1, <i8 165, i8 165, i8 165, i8 165, i8 165, i8 165, i8 165, i8 165, i8 165, i8 165, i8 165, i8 165, i8 165, i8 165, i8 165, i8 165>
%vshl_n = shl <16 x i8> %src2, <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1>
%result = or <16 x i8> %and.i, %vshl_n
store <16 x i8> %result, <16 x i8>* %dest, align 16
ret void
}
define void @testRightGood(<16 x i8> %src1, <16 x i8> %src2, <16 x i8>* %dest) nounwind {
; CHECK-LABEL: testRightGood:
; CHECK: sri.16b v0, v1, #3
%and.i = and <16 x i8> %src1, <i8 252, i8 252, i8 252, i8 252, i8 252, i8 252, i8 252, i8 252, i8 252, i8 252, i8 252, i8 252, i8 252, i8 252, i8 252, i8 252>
%vshl_n = lshr <16 x i8> %src2, <i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3>
%result = or <16 x i8> %and.i, %vshl_n
store <16 x i8> %result, <16 x i8>* %dest, align 16
ret void
}
define void @testRightBad(<16 x i8> %src1, <16 x i8> %src2, <16 x i8>* %dest) nounwind {
; CHECK-LABEL: testRightBad:
; CHECK-NOT: sri
%and.i = and <16 x i8> %src1, <i8 165, i8 165, i8 165, i8 165, i8 165, i8 165, i8 165, i8 165, i8 165, i8 165, i8 165, i8 165, i8 165, i8 165, i8 165, i8 165>
%vshl_n = lshr <16 x i8> %src2, <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1>
%result = or <16 x i8> %and.i, %vshl_n
store <16 x i8> %result, <16 x i8>* %dest, align 16
ret void
}