x86-f16c.ll
2.95 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
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
; RUN: opt < %s -instcombine -S | FileCheck %s
declare <4 x float> @llvm.x86.vcvtph2ps.128(<8 x i16>)
declare <8 x float> @llvm.x86.vcvtph2ps.256(<8 x i16>)
;
; Vector Demanded Elts
;
; Only bottom 4 elements required.
define <4 x float> @demand_vcvtph2ps_128(<8 x i16> %A) {
; CHECK-LABEL: @demand_vcvtph2ps_128(
; CHECK-NEXT: [[TMP1:%.*]] = shufflevector <8 x i16> [[A:%.*]], <8 x i16> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
; CHECK-NEXT: [[TMP2:%.*]] = bitcast <4 x i16> [[TMP1]] to <4 x half>
; CHECK-NEXT: [[CVTPH2PS:%.*]] = fpext <4 x half> [[TMP2]] to <4 x float>
; CHECK-NEXT: ret <4 x float> [[CVTPH2PS]]
;
%1 = shufflevector <8 x i16> %A, <8 x i16> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3>
%2 = tail call <4 x float> @llvm.x86.vcvtph2ps.128(<8 x i16> %1)
ret <4 x float> %2
}
; All 8 elements required.
define <8 x float> @demand_vcvtph2ps_256(<8 x i16> %A) {
; CHECK-LABEL: @demand_vcvtph2ps_256(
; CHECK-NEXT: [[TMP1:%.*]] = shufflevector <8 x i16> [[A:%.*]], <8 x i16> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3>
; CHECK-NEXT: [[TMP2:%.*]] = bitcast <8 x i16> [[TMP1]] to <8 x half>
; CHECK-NEXT: [[CVTPH2PS:%.*]] = fpext <8 x half> [[TMP2]] to <8 x float>
; CHECK-NEXT: ret <8 x float> [[CVTPH2PS]]
;
%1 = shufflevector <8 x i16> %A, <8 x i16> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3>
%2 = tail call <8 x float> @llvm.x86.vcvtph2ps.256(<8 x i16> %1)
ret <8 x float> %2
}
;
; Constant Folding
;
define <4 x float> @fold_vcvtph2ps_128() {
; CHECK-LABEL: @fold_vcvtph2ps_128(
; CHECK-NEXT: ret <4 x float> <float 0.000000e+00, float 5.000000e-01, float 1.000000e+00, float -0.000000e+00>
;
%1 = tail call <4 x float> @llvm.x86.vcvtph2ps.128(<8 x i16> <i16 0, i16 14336, i16 15360, i16 32768, i16 16384, i16 31743, i16 48128, i16 49152>)
ret <4 x float> %1
}
define <8 x float> @fold_vcvtph2ps_256() {
; CHECK-LABEL: @fold_vcvtph2ps_256(
; CHECK-NEXT: ret <8 x float> <float 0.000000e+00, float 5.000000e-01, float 1.000000e+00, float -0.000000e+00, float 2.000000e+00, float 6.550400e+04, float -1.000000e+00, float -2.000000e+00>
;
%1 = tail call <8 x float> @llvm.x86.vcvtph2ps.256(<8 x i16> <i16 0, i16 14336, i16 15360, i16 32768, i16 16384, i16 31743, i16 48128, i16 49152>)
ret <8 x float> %1
}
define <4 x float> @fold_vcvtph2ps_128_zero() {
; CHECK-LABEL: @fold_vcvtph2ps_128_zero(
; CHECK-NEXT: ret <4 x float> zeroinitializer
;
%1 = tail call <4 x float> @llvm.x86.vcvtph2ps.128(<8 x i16> <i16 0, i16 0, i16 0, i16 0, i16 0, i16 0, i16 0, i16 0>)
ret <4 x float> %1
}
define <8 x float> @fold_vcvtph2ps_256_zero() {
; CHECK-LABEL: @fold_vcvtph2ps_256_zero(
; CHECK-NEXT: ret <8 x float> zeroinitializer
;
%1 = tail call <8 x float> @llvm.x86.vcvtph2ps.256(<8 x i16> <i16 0, i16 0, i16 0, i16 0, i16 0, i16 0, i16 0, i16 0>)
ret <8 x float> %1
}