-
Notifications
You must be signed in to change notification settings - Fork 10.7k
/
mma-acc-spill.ll
122 lines (121 loc) · 5.05 KB
/
mma-acc-spill.ll
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
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc -verify-machineinstrs -mtriple=powerpc64le-unknown-linux-gnu \
; RUN: -mcpu=pwr10 -ppc-asm-full-reg-names \
; RUN: -ppc-vsr-nums-as-vr < %s | FileCheck %s
; RUN: llc -verify-machineinstrs -mtriple=powerpc64-unknown-linux-gnu \
; RUN: -mcpu=pwr10 -ppc-asm-full-reg-names \
; RUN: -ppc-vsr-nums-as-vr < %s | FileCheck %s --check-prefix=CHECK-BE
declare <512 x i1> @llvm.ppc.mma.xvf16ger2pp(<512 x i1>, <16 x i8>, <16 x i8>)
declare <512 x i1> @llvm.ppc.mma.assemble.acc(<16 x i8>, <16 x i8>, <16 x i8>, <16 x i8>)
declare void @foo()
define void @intrinsics1(<16 x i8> %vc1, <16 x i8> %vc2, <16 x i8> %vc3, <16 x i8> %vc4, i8* %ptr) {
; CHECK-LABEL: intrinsics1:
; CHECK: # %bb.0:
; CHECK-NEXT: mflr r0
; CHECK-NEXT: std r0, 16(r1)
; CHECK-NEXT: stdu r1, -176(r1)
; CHECK-NEXT: .cfi_def_cfa_offset 176
; CHECK-NEXT: .cfi_offset lr, 16
; CHECK-NEXT: .cfi_offset r30, -16
; CHECK-NEXT: .cfi_offset v28, -80
; CHECK-NEXT: .cfi_offset v29, -64
; CHECK-NEXT: .cfi_offset v30, -48
; CHECK-NEXT: .cfi_offset v31, -32
; CHECK-NEXT: stxv v28, 96(r1) # 16-byte Folded Spill
; CHECK-NEXT: stxv v29, 112(r1) # 16-byte Folded Spill
; CHECK-NEXT: vmr v29, v3
; CHECK-NEXT: vmr v28, v2
; CHECK-NEXT: xxlor vs0, v28, v28
; CHECK-NEXT: stxv v30, 128(r1) # 16-byte Folded Spill
; CHECK-NEXT: stxv v31, 144(r1) # 16-byte Folded Spill
; CHECK-NEXT: vmr v31, v5
; CHECK-NEXT: vmr v30, v4
; CHECK-NEXT: std r30, 160(r1) # 8-byte Folded Spill
; CHECK-NEXT: xxlor vs1, v29, v29
; CHECK-NEXT: xxlor vs2, v30, v30
; CHECK-NEXT: xxlor vs3, v31, v31
; CHECK-NEXT: ld r30, 272(r1)
; CHECK-NEXT: xxmtacc acc0
; CHECK-NEXT: xvf16ger2pp acc0, v2, v4
; CHECK-NEXT: xxmfacc acc0
; CHECK-NEXT: stxvp vsp0, 64(r1)
; CHECK-NEXT: stxvp vsp2, 32(r1)
; CHECK-NEXT: bl foo@notoc
; CHECK-NEXT: lxvp vsp0, 64(r1)
; CHECK-NEXT: lxvp vsp2, 32(r1)
; CHECK-NEXT: xxmtacc acc0
; CHECK-NEXT: xvf16ger2pp acc0, v28, v30
; CHECK-NEXT: lxv v31, 144(r1) # 16-byte Folded Reload
; CHECK-NEXT: lxv v30, 128(r1) # 16-byte Folded Reload
; CHECK-NEXT: lxv v29, 112(r1) # 16-byte Folded Reload
; CHECK-NEXT: lxv v28, 96(r1) # 16-byte Folded Reload
; CHECK-NEXT: xxmfacc acc0
; CHECK-NEXT: stxv vs0, 48(r30)
; CHECK-NEXT: stxv vs1, 32(r30)
; CHECK-NEXT: stxv vs2, 16(r30)
; CHECK-NEXT: stxv vs3, 0(r30)
; CHECK-NEXT: ld r30, 160(r1) # 8-byte Folded Reload
; CHECK-NEXT: addi r1, r1, 176
; CHECK-NEXT: ld r0, 16(r1)
; CHECK-NEXT: mtlr r0
; CHECK-NEXT: blr
;
; CHECK-BE-LABEL: intrinsics1:
; CHECK-BE: # %bb.0:
; CHECK-BE-NEXT: mflr r0
; CHECK-BE-NEXT: std r0, 16(r1)
; CHECK-BE-NEXT: stdu r1, -256(r1)
; CHECK-BE-NEXT: .cfi_def_cfa_offset 256
; CHECK-BE-NEXT: .cfi_offset lr, 16
; CHECK-BE-NEXT: .cfi_offset r30, -16
; CHECK-BE-NEXT: .cfi_offset v28, -80
; CHECK-BE-NEXT: .cfi_offset v29, -64
; CHECK-BE-NEXT: .cfi_offset v30, -48
; CHECK-BE-NEXT: .cfi_offset v31, -32
; CHECK-BE-NEXT: stxv v28, 176(r1) # 16-byte Folded Spill
; CHECK-BE-NEXT: stxv v29, 192(r1) # 16-byte Folded Spill
; CHECK-BE-NEXT: vmr v29, v3
; CHECK-BE-NEXT: vmr v28, v2
; CHECK-BE-NEXT: xxlor vs0, v28, v28
; CHECK-BE-NEXT: stxv v30, 208(r1) # 16-byte Folded Spill
; CHECK-BE-NEXT: stxv v31, 224(r1) # 16-byte Folded Spill
; CHECK-BE-NEXT: vmr v31, v5
; CHECK-BE-NEXT: vmr v30, v4
; CHECK-BE-NEXT: std r30, 240(r1) # 8-byte Folded Spill
; CHECK-BE-NEXT: xxlor vs1, v29, v29
; CHECK-BE-NEXT: xxlor vs2, v30, v30
; CHECK-BE-NEXT: xxlor vs3, v31, v31
; CHECK-BE-NEXT: ld r30, 368(r1)
; CHECK-BE-NEXT: xxmtacc acc0
; CHECK-BE-NEXT: xvf16ger2pp acc0, v2, v4
; CHECK-BE-NEXT: xxmfacc acc0
; CHECK-BE-NEXT: stxvp vsp0, 112(r1)
; CHECK-BE-NEXT: stxvp vsp2, 144(r1)
; CHECK-BE-NEXT: bl foo
; CHECK-BE-NEXT: nop
; CHECK-BE-NEXT: lxvp vsp0, 112(r1)
; CHECK-BE-NEXT: lxvp vsp2, 144(r1)
; CHECK-BE-NEXT: xxmtacc acc0
; CHECK-BE-NEXT: xvf16ger2pp acc0, v28, v30
; CHECK-BE-NEXT: lxv v31, 224(r1) # 16-byte Folded Reload
; CHECK-BE-NEXT: lxv v30, 208(r1) # 16-byte Folded Reload
; CHECK-BE-NEXT: lxv v29, 192(r1) # 16-byte Folded Reload
; CHECK-BE-NEXT: lxv v28, 176(r1) # 16-byte Folded Reload
; CHECK-BE-NEXT: xxmfacc acc0
; CHECK-BE-NEXT: stxv vs1, 16(r30)
; CHECK-BE-NEXT: stxv vs0, 0(r30)
; CHECK-BE-NEXT: stxv vs3, 48(r30)
; CHECK-BE-NEXT: stxv vs2, 32(r30)
; CHECK-BE-NEXT: ld r30, 240(r1) # 8-byte Folded Reload
; CHECK-BE-NEXT: addi r1, r1, 256
; CHECK-BE-NEXT: ld r0, 16(r1)
; CHECK-BE-NEXT: mtlr r0
; CHECK-BE-NEXT: blr
%1 = tail call <512 x i1> @llvm.ppc.mma.assemble.acc(<16 x i8> %vc1, <16 x i8> %vc2, <16 x i8> %vc3, <16 x i8> %vc4)
%2 = tail call <512 x i1> @llvm.ppc.mma.xvf16ger2pp(<512 x i1> %1, <16 x i8> %vc1, <16 x i8> %vc3)
tail call void @foo()
%3 = tail call <512 x i1> @llvm.ppc.mma.xvf16ger2pp(<512 x i1> %2, <16 x i8> %vc1, <16 x i8> %vc3)
%4 = bitcast i8* %ptr to <512 x i1>*
store <512 x i1> %3, <512 x i1>* %4, align 64
ret void
}