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
|
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc -mtriple=aarch64-linux-gnu -mattr=+sve2p1 -mattr=+sve-b16b16 -mattr=+use-experimental-zeroing-pseudos -verify-machineinstrs < %s \
; RUN: | FileCheck %s
define <vscale x 8 x bfloat> @bfmax_pred(<vscale x 8 x i1> %pg, <vscale x 8 x bfloat> %a, <vscale x 8 x bfloat> %b){
; CHECK-LABEL: bfmax_pred:
; CHECK: // %bb.0:
; CHECK-NEXT: bfmax z0.h, p0/m, z0.h, z1.h
; CHECK-NEXT: ret
%res = call <vscale x 8 x bfloat> @llvm.aarch64.sve.fmax.nxv8bf16(<vscale x 8 x i1> %pg, <vscale x 8 x bfloat> %a, <vscale x 8 x bfloat> %b)
ret <vscale x 8 x bfloat> %res
}
define <vscale x 8 x bfloat> @bfmax(<vscale x 8 x bfloat> %a, <vscale x 8 x bfloat> %b){
; CHECK-LABEL: bfmax:
; CHECK: // %bb.0:
; CHECK-NEXT: ptrue p0.h
; CHECK-NEXT: bfmax z0.h, p0/m, z0.h, z1.h
; CHECK-NEXT: ret
%elt = call <vscale x 8 x i1> @llvm.aarch64.sve.ptrue.nxv8i1(i32 31)
%res = call <vscale x 8 x bfloat> @llvm.aarch64.sve.fmax.nxv8bf16(<vscale x 8 x i1> %elt, <vscale x 8 x bfloat> %a, <vscale x 8 x bfloat> %b)
ret <vscale x 8 x bfloat> %res
}
define <vscale x 8 x bfloat> @bfmax_zeroing(<vscale x 8 x i1> %pg, <vscale x 8 x bfloat> %a, <vscale x 8 x bfloat> %b) {
; CHECK-LABEL: bfmax_zeroing:
; CHECK: // %bb.0:
; CHECK-NEXT: movprfx z0.h, p0/z, z0.h
; CHECK-NEXT: bfmax z0.h, p0/m, z0.h, z1.h
; CHECK-NEXT: ret
%a_z = select <vscale x 8 x i1> %pg, <vscale x 8 x bfloat> %a, <vscale x 8 x bfloat> zeroinitializer
%out = call <vscale x 8 x bfloat> @llvm.aarch64.sve.fmax.nxv8bf16(<vscale x 8 x i1> %pg,
<vscale x 8 x bfloat> %a_z,
<vscale x 8 x bfloat> %b)
ret <vscale x 8 x bfloat> %out
}
define <vscale x 8 x bfloat> @bfmax_u_pred(<vscale x 8 x i1> %pg, <vscale x 8 x bfloat> %a, <vscale x 8 x bfloat> %b){
; CHECK-LABEL: bfmax_u_pred:
; CHECK: // %bb.0:
; CHECK-NEXT: bfmax z0.h, p0/m, z0.h, z1.h
; CHECK-NEXT: ret
%res = call <vscale x 8 x bfloat> @llvm.aarch64.sve.fmax.u.nxv8bf16(<vscale x 8 x i1> %pg, <vscale x 8 x bfloat> %a, <vscale x 8 x bfloat> %b)
ret <vscale x 8 x bfloat> %res
}
define <vscale x 8 x bfloat> @bfmax_u(<vscale x 8 x bfloat> %a, <vscale x 8 x bfloat> %b){
; CHECK-LABEL: bfmax_u:
; CHECK: // %bb.0:
; CHECK-NEXT: ptrue p0.h
; CHECK-NEXT: bfmax z0.h, p0/m, z0.h, z1.h
; CHECK-NEXT: ret
%elt = call <vscale x 8 x i1> @llvm.aarch64.sve.ptrue.nxv8i1(i32 31)
%res = call <vscale x 8 x bfloat> @llvm.aarch64.sve.fmax.u.nxv8bf16(<vscale x 8 x i1> %elt, <vscale x 8 x bfloat> %a, <vscale x 8 x bfloat> %b)
ret <vscale x 8 x bfloat> %res
}
define <vscale x 8 x bfloat> @bfmax_u_zeroing(<vscale x 8 x i1> %pg, <vscale x 8 x bfloat> %a, <vscale x 8 x bfloat> %b) {
; CHECK-LABEL: bfmax_u_zeroing:
; CHECK: // %bb.0:
; CHECK-NEXT: movi v2.2d, #0000000000000000
; CHECK-NEXT: sel z0.h, p0, z0.h, z2.h
; CHECK-NEXT: bfmax z0.h, p0/m, z0.h, z1.h
; CHECK-NEXT: ret
%a_z = select <vscale x 8 x i1> %pg, <vscale x 8 x bfloat> %a, <vscale x 8 x bfloat> zeroinitializer
%out = call <vscale x 8 x bfloat> @llvm.aarch64.sve.fmax.u.nxv8bf16(<vscale x 8 x i1> %pg,
<vscale x 8 x bfloat> %a_z,
<vscale x 8 x bfloat> %b)
ret <vscale x 8 x bfloat> %out
}
declare <vscale x 8 x bfloat> @llvm.aarch64.sve.fmax.nxv8bf16(<vscale x 8 x i1>, <vscale x 8 x bfloat>, <vscale x 8 x bfloat>)
declare <vscale x 8 x bfloat> @llvm.aarch64.sve.fmax.u.nxv8bf16(<vscale x 8 x i1>, <vscale x 8 x bfloat>, <vscale x 8 x bfloat>)
declare <vscale x 8 x i1> @llvm.aarch64.sve.ptrue.nxv8i1(i32 immarg)
|
