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
72
73
74
75
76
77
78
|
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 5
; RUN: llc -mattr=+sve2,+fp8 < %s | FileCheck %s
; RUN: llc -mattr=+sme2,+fp8 --force-streaming < %s | FileCheck %s
target triple = "aarch64-linux"
define <vscale x 8 x bfloat> @cvt1_bf16(<vscale x 16 x i8> %s) {
; CHECK-LABEL: cvt1_bf16:
; CHECK: // %bb.0:
; CHECK-NEXT: bf1cvt z0.h, z0.b
; CHECK-NEXT: ret
%r = call <vscale x 8 x bfloat> @llvm.aarch64.sve.fp8.cvt1.nxv8bf16(<vscale x 16 x i8> %s)
ret <vscale x 8 x bfloat> %r
}
define <vscale x 8 x bfloat> @cvt2_bf16(<vscale x 16 x i8> %s) {
; CHECK-LABEL: cvt2_bf16:
; CHECK: // %bb.0:
; CHECK-NEXT: bf2cvt z0.h, z0.b
; CHECK-NEXT: ret
%r = call <vscale x 8 x bfloat> @llvm.aarch64.sve.fp8.cvt2.nxv8bf16(<vscale x 16 x i8> %s)
ret <vscale x 8 x bfloat> %r
}
define <vscale x 8 x bfloat> @cvtlt1_bf16(<vscale x 16 x i8> %s) {
; CHECK-LABEL: cvtlt1_bf16:
; CHECK: // %bb.0:
; CHECK-NEXT: bf1cvtlt z0.h, z0.b
; CHECK-NEXT: ret
%r = call <vscale x 8 x bfloat> @llvm.aarch64.sve.fp8.cvtlt1.nxv8bf16(<vscale x 16 x i8> %s)
ret <vscale x 8 x bfloat> %r
}
define <vscale x 8 x bfloat> @cvtlt2_bf16(<vscale x 16 x i8> %s) {
; CHECK-LABEL: cvtlt2_bf16:
; CHECK: // %bb.0:
; CHECK-NEXT: bf2cvtlt z0.h, z0.b
; CHECK-NEXT: ret
%r = call <vscale x 8 x bfloat> @llvm.aarch64.sve.fp8.cvtlt2.nxv8bf16(<vscale x 16 x i8> %s)
ret <vscale x 8 x bfloat> %r
}
define <vscale x 8 x half> @cvt1_f16(<vscale x 16 x i8> %s) {
; CHECK-LABEL: cvt1_f16:
; CHECK: // %bb.0:
; CHECK-NEXT: f1cvt z0.h, z0.b
; CHECK-NEXT: ret
%r = call <vscale x 8 x half> @llvm.aarch64.sve.fp8.cvt1.nxv8f16(<vscale x 16 x i8> %s)
ret <vscale x 8 x half> %r
}
define <vscale x 8 x half> @cvt2_f16(<vscale x 16 x i8> %s) {
; CHECK-LABEL: cvt2_f16:
; CHECK: // %bb.0:
; CHECK-NEXT: f2cvt z0.h, z0.b
; CHECK-NEXT: ret
%r = call <vscale x 8 x half> @llvm.aarch64.sve.fp8.cvt2.nxv8f16(<vscale x 16 x i8> %s)
ret <vscale x 8 x half> %r
}
define <vscale x 8 x half> @cvtlt1_f16(<vscale x 16 x i8> %s) {
; CHECK-LABEL: cvtlt1_f16:
; CHECK: // %bb.0:
; CHECK-NEXT: f1cvtlt z0.h, z0.b
; CHECK-NEXT: ret
%r = call <vscale x 8 x half> @llvm.aarch64.sve.fp8.cvtlt1.nxv8f16(<vscale x 16 x i8> %s)
ret <vscale x 8 x half> %r
}
define <vscale x 8 x half> @cvtlt2_f16(<vscale x 16 x i8> %s) {
; CHECK-LABEL: cvtlt2_f16:
; CHECK: // %bb.0:
; CHECK-NEXT: f2cvtlt z0.h, z0.b
; CHECK-NEXT: ret
%r = call <vscale x 8 x half> @llvm.aarch64.sve.fp8.cvtlt2.nxv8f16(<vscale x 16 x i8> %s)
ret <vscale x 8 x half> %r
}
|
