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
|
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc -mtriple=aarch64-linux-gnu -mattr=+sve2 < %s | FileCheck %s
;
; LDNT1B, LDNT1W, LDNT1H, LDNT1D: base + 32-bit unscaled offsets, zero (uxtw)
; extended to 64 bits.
; e.g. ldnt1h { z0.s }, p0/z, [z0.s, x0]
;
; LDNT1B
define <vscale x 4 x i32> @gldnt1b_s_uxtw(<vscale x 4 x i1> %pg, ptr %base, <vscale x 4 x i32> %b) {
; CHECK-LABEL: gldnt1b_s_uxtw:
; CHECK: // %bb.0:
; CHECK-NEXT: ldnt1b { z0.s }, p0/z, [z0.s, x0]
; CHECK-NEXT: ret
%load = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnt1.gather.uxtw.nxv4i8(<vscale x 4 x i1> %pg,
ptr %base,
<vscale x 4 x i32> %b)
%res = zext <vscale x 4 x i8> %load to <vscale x 4 x i32>
ret <vscale x 4 x i32> %res
}
; LDNT1H
define <vscale x 4 x i32> @gldnt1h_s_uxtw(<vscale x 4 x i1> %pg, ptr %base, <vscale x 4 x i32> %b) {
; CHECK-LABEL: gldnt1h_s_uxtw:
; CHECK: // %bb.0:
; CHECK-NEXT: ldnt1h { z0.s }, p0/z, [z0.s, x0]
; CHECK-NEXT: ret
%load = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnt1.gather.uxtw.nxv4i16(<vscale x 4 x i1> %pg,
ptr %base,
<vscale x 4 x i32> %b)
%res = zext <vscale x 4 x i16> %load to <vscale x 4 x i32>
ret <vscale x 4 x i32> %res
}
; LDNT1W
define <vscale x 4 x i32> @gldnt1w_s_uxtw(<vscale x 4 x i1> %pg, ptr %base, <vscale x 4 x i32> %b) {
; CHECK-LABEL: gldnt1w_s_uxtw:
; CHECK: // %bb.0:
; CHECK-NEXT: ldnt1w { z0.s }, p0/z, [z0.s, x0]
; CHECK-NEXT: ret
%load = call <vscale x 4 x i32> @llvm.aarch64.sve.ldnt1.gather.uxtw.nxv4i32(<vscale x 4 x i1> %pg,
ptr %base,
<vscale x 4 x i32> %b)
ret <vscale x 4 x i32> %load
}
define <vscale x 4 x float> @gldnt1w_s_uxtw_float(<vscale x 4 x i1> %pg, ptr %base, <vscale x 4 x i32> %b) {
; CHECK-LABEL: gldnt1w_s_uxtw_float:
; CHECK: // %bb.0:
; CHECK-NEXT: ldnt1w { z0.s }, p0/z, [z0.s, x0]
; CHECK-NEXT: ret
%load = call <vscale x 4 x float> @llvm.aarch64.sve.ldnt1.gather.uxtw.nxv4f32(<vscale x 4 x i1> %pg,
ptr %base,
<vscale x 4 x i32> %b)
ret <vscale x 4 x float> %load
}
; LDNT1SB, LDNT1SW, LDNT1SH: base + 32-bit unscaled offsets, zero (uxtw)
; extended to 64 bits.
; e.g. ldnt1sh { z0.s }, p0/z, [z0.s, x0]
;
; LDNT1SB
define <vscale x 4 x i32> @gldnt1sb_s_uxtw(<vscale x 4 x i1> %pg, ptr %base, <vscale x 4 x i32> %b) {
; CHECK-LABEL: gldnt1sb_s_uxtw:
; CHECK: // %bb.0:
; CHECK-NEXT: ldnt1sb { z0.s }, p0/z, [z0.s, x0]
; CHECK-NEXT: ret
%load = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnt1.gather.uxtw.nxv4i8(<vscale x 4 x i1> %pg,
ptr %base,
<vscale x 4 x i32> %b)
%res = sext <vscale x 4 x i8> %load to <vscale x 4 x i32>
ret <vscale x 4 x i32> %res
}
; LDNT1SH
define <vscale x 4 x i32> @gldnt1sh_s_uxtw(<vscale x 4 x i1> %pg, ptr %base, <vscale x 4 x i32> %b) {
; CHECK-LABEL: gldnt1sh_s_uxtw:
; CHECK: // %bb.0:
; CHECK-NEXT: ldnt1sh { z0.s }, p0/z, [z0.s, x0]
; CHECK-NEXT: ret
%load = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnt1.gather.uxtw.nxv4i16(<vscale x 4 x i1> %pg,
ptr %base,
<vscale x 4 x i32> %b)
%res = sext <vscale x 4 x i16> %load to <vscale x 4 x i32>
ret <vscale x 4 x i32> %res
}
; LDNT1B/LDNT1SB
declare <vscale x 4 x i8> @llvm.aarch64.sve.ldnt1.gather.uxtw.nxv4i8(<vscale x 4 x i1>, ptr, <vscale x 4 x i32>)
declare <vscale x 4 x i8> @llvm.aarch64.sve.ldnt1.gather.sxtw.nxv4i8(<vscale x 4 x i1>, ptr, <vscale x 4 x i32>)
; LDNT1H/LDNT1SH
declare <vscale x 4 x i16> @llvm.aarch64.sve.ldnt1.gather.sxtw.nxv4i16(<vscale x 4 x i1>, ptr, <vscale x 4 x i32>)
declare <vscale x 4 x i16> @llvm.aarch64.sve.ldnt1.gather.uxtw.nxv4i16(<vscale x 4 x i1>, ptr, <vscale x 4 x i32>)
; LDNT1W/LDNT1SW
declare <vscale x 4 x i32> @llvm.aarch64.sve.ldnt1.gather.sxtw.nxv4i32(<vscale x 4 x i1>, ptr, <vscale x 4 x i32>)
declare <vscale x 4 x i32> @llvm.aarch64.sve.ldnt1.gather.uxtw.nxv4i32(<vscale x 4 x i1>, ptr, <vscale x 4 x i32>)
declare <vscale x 4 x float> @llvm.aarch64.sve.ldnt1.gather.sxtw.nxv4f32(<vscale x 4 x i1>, ptr, <vscale x 4 x i32>)
declare <vscale x 4 x float> @llvm.aarch64.sve.ldnt1.gather.uxtw.nxv4f32(<vscale x 4 x i1>, ptr, <vscale x 4 x i32>)
|
