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
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
|
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 5
; RUN: llc -force-streaming -verify-machineinstrs < %s | FileCheck %s
target triple = "aarch64-linux"
define void @mop4a_za16_fp8_1x1(<vscale x 16 x i8> %zn, <vscale x 16 x i8> %zm) #0 {
; CHECK-LABEL: mop4a_za16_fp8_1x1:
; CHECK: // %bb.0:
; CHECK-NEXT: mov z24.d, z1.d
; CHECK-NEXT: fmop4a za0.h, z0.b, z24.b
; CHECK-NEXT: ret
call void @llvm.aarch64.sme.fp8.fmop4a.za16.1x1(i32 0, <vscale x 16 x i8> %zn, <vscale x 16 x i8> %zm)
ret void
}
define void @mop4a_za16_fp8_1x2(<vscale x 16 x i8> %zn, <vscale x 16 x i8> %zm1, <vscale x 16 x i8> %zm2) #0 {
; CHECK-LABEL: mop4a_za16_fp8_1x2:
; CHECK: // %bb.0:
; CHECK-NEXT: mov z25.d, z2.d
; CHECK-NEXT: mov z24.d, z1.d
; CHECK-NEXT: fmop4a za0.h, z0.b, { z24.b, z25.b }
; CHECK-NEXT: ret
call void @llvm.aarch64.sme.fp8.fmop4a.za16.1x2(i32 0, <vscale x 16 x i8> %zn, <vscale x 16 x i8> %zm1, <vscale x 16 x i8> %zm2)
ret void
}
define void @mop4a_za16_fp8_2x1(<vscale x 16 x i8> %zn1, <vscale x 16 x i8> %zn2, <vscale x 16 x i8> %zm) #0 {
; CHECK-LABEL: mop4a_za16_fp8_2x1:
; CHECK: // %bb.0:
; CHECK-NEXT: mov z24.d, z2.d
; CHECK-NEXT: // kill: def $z1 killed $z1 killed $z0_z1 def $z0_z1
; CHECK-NEXT: // kill: def $z0 killed $z0 killed $z0_z1 def $z0_z1
; CHECK-NEXT: fmop4a za0.h, { z0.b, z1.b }, z24.b
; CHECK-NEXT: ret
call void @llvm.aarch64.sme.fp8.fmop4a.za16.2x1(i32 0, <vscale x 16 x i8> %zn1, <vscale x 16 x i8> %zn2, <vscale x 16 x i8> %zm)
ret void
}
define void @mop4a_za16_fp8_2x2(<vscale x 16 x i8> %zn1, <vscale x 16 x i8> %zn2, <vscale x 16 x i8> %zm1, <vscale x 16 x i8> %zm2) #0 {
; CHECK-LABEL: mop4a_za16_fp8_2x2:
; CHECK: // %bb.0:
; CHECK-NEXT: mov z25.d, z3.d
; CHECK-NEXT: // kill: def $z1 killed $z1 killed $z0_z1 def $z0_z1
; CHECK-NEXT: mov z24.d, z2.d
; CHECK-NEXT: // kill: def $z0 killed $z0 killed $z0_z1 def $z0_z1
; CHECK-NEXT: fmop4a za0.h, { z0.b, z1.b }, { z24.b, z25.b }
; CHECK-NEXT: ret
call void @llvm.aarch64.sme.fp8.fmop4a.za16.2x2(i32 0, <vscale x 16 x i8> %zn1, <vscale x 16 x i8> %zn2, <vscale x 16 x i8> %zm1, <vscale x 16 x i8> %zm2)
ret void
}
define void @mop4a_za32_fp8(<vscale x 16 x i8> %zn, <vscale x 16 x i8> %zm) #0 {
; CHECK-LABEL: mop4a_za32_fp8:
; CHECK: // %bb.0:
; CHECK-NEXT: mov z24.d, z1.d
; CHECK-NEXT: fmop4a za0.s, z0.b, z24.b
; CHECK-NEXT: ret
call void @llvm.aarch64.sme.fp8.fmop4a.za32.1x1(i32 0, <vscale x 16 x i8> %zn, <vscale x 16 x i8> %zm)
ret void
}
define void @mop4a_za32_fp8_1x2(<vscale x 16 x i8> %zn, <vscale x 16 x i8> %zm1, <vscale x 16 x i8> %zm2) #0 {
; CHECK-LABEL: mop4a_za32_fp8_1x2:
; CHECK: // %bb.0:
; CHECK-NEXT: mov z25.d, z2.d
; CHECK-NEXT: mov z24.d, z1.d
; CHECK-NEXT: fmop4a za0.s, z0.b, { z24.b, z25.b }
; CHECK-NEXT: ret
call void @llvm.aarch64.sme.fp8.fmop4a.za32.1x2(i32 0, <vscale x 16 x i8> %zn, <vscale x 16 x i8> %zm1, <vscale x 16 x i8> %zm2)
ret void
}
define void @mop4a_za32_fp8_2x1(<vscale x 16 x i8> %zn1, <vscale x 16 x i8> %zn2, <vscale x 16 x i8> %zm) #0 {
; CHECK-LABEL: mop4a_za32_fp8_2x1:
; CHECK: // %bb.0:
; CHECK-NEXT: mov z24.d, z2.d
; CHECK-NEXT: // kill: def $z1 killed $z1 killed $z0_z1 def $z0_z1
; CHECK-NEXT: // kill: def $z0 killed $z0 killed $z0_z1 def $z0_z1
; CHECK-NEXT: fmop4a za0.s, { z0.b, z1.b }, z24.b
; CHECK-NEXT: ret
call void @llvm.aarch64.sme.fp8.fmop4a.za32.2x1(i32 0, <vscale x 16 x i8> %zn1, <vscale x 16 x i8> %zn2, <vscale x 16 x i8> %zm)
ret void
}
define void @mop4a_za32_fp8_2x2(<vscale x 16 x i8> %zn1, <vscale x 16 x i8> %zn2, <vscale x 16 x i8> %zm1, <vscale x 16 x i8> %zm2) #0 {
; CHECK-LABEL: mop4a_za32_fp8_2x2:
; CHECK: // %bb.0:
; CHECK-NEXT: mov z25.d, z3.d
; CHECK-NEXT: // kill: def $z1 killed $z1 killed $z0_z1 def $z0_z1
; CHECK-NEXT: mov z24.d, z2.d
; CHECK-NEXT: // kill: def $z0 killed $z0 killed $z0_z1 def $z0_z1
; CHECK-NEXT: fmop4a za0.s, { z0.b, z1.b }, { z24.b, z25.b }
; CHECK-NEXT: ret
call void @llvm.aarch64.sme.fp8.fmop4a.za32.2x2(i32 0, <vscale x 16 x i8> %zn1, <vscale x 16 x i8> %zn2, <vscale x 16 x i8> %zm1, <vscale x 16 x i8> %zm2)
ret void
}
attributes #0 = {nounwind "target-features" = "+sme-f8f16,+sme-f8f32,+sme2p1,+sme-mop4" }
|
