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
|
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc -mtriple=aarch64-linux-gnu -mattr=+sve2 < %s | FileCheck %s
; RUN: llc -mtriple=aarch64-linux-gnu -mattr=+sme -force-streaming < %s | FileCheck %s
;
; CDOT
;
define <vscale x 4 x i32> @cdot_s(<vscale x 4 x i32> %a, <vscale x 16 x i8> %b, <vscale x 16 x i8> %c) {
; CHECK-LABEL: cdot_s:
; CHECK: // %bb.0:
; CHECK-NEXT: cdot z0.s, z1.b, z2.b, #0
; CHECK-NEXT: ret
%out = call <vscale x 4 x i32> @llvm.aarch64.sve.cdot.nxv4i32(<vscale x 4 x i32> %a,
<vscale x 16 x i8> %b,
<vscale x 16 x i8> %c,
i32 0)
ret <vscale x 4 x i32> %out
}
define <vscale x 2 x i64> @cdot_d(<vscale x 2 x i64> %a, <vscale x 8 x i16> %b, <vscale x 8 x i16> %c) {
; CHECK-LABEL: cdot_d:
; CHECK: // %bb.0:
; CHECK-NEXT: cdot z0.d, z1.h, z2.h, #90
; CHECK-NEXT: ret
%out = call <vscale x 2 x i64> @llvm.aarch64.sve.cdot.nxv2i64(<vscale x 2 x i64> %a,
<vscale x 8 x i16> %b,
<vscale x 8 x i16> %c,
i32 90)
ret <vscale x 2 x i64> %out
}
;
; CDOT(indexed)
;
define <vscale x 4 x i32> @cdot_s_idx(<vscale x 4 x i32> %a, <vscale x 16 x i8> %b, <vscale x 16 x i8> %c) {
; CHECK-LABEL: cdot_s_idx:
; CHECK: // %bb.0:
; CHECK-NEXT: cdot z0.s, z1.b, z2.b[0], #180
; CHECK-NEXT: ret
%out = call <vscale x 4 x i32> @llvm.aarch64.sve.cdot.lane.nxv4i32(<vscale x 4 x i32> %a,
<vscale x 16 x i8> %b,
<vscale x 16 x i8> %c,
i32 0, i32 180)
ret <vscale x 4 x i32> %out
}
define <vscale x 2 x i64> @cdot_d_idx(<vscale x 2 x i64> %a, <vscale x 8 x i16> %b, <vscale x 8 x i16> %c) {
; CHECK-LABEL: cdot_d_idx:
; CHECK: // %bb.0:
; CHECK-NEXT: cdot z0.d, z1.h, z2.h[1], #270
; CHECK-NEXT: ret
%out = call <vscale x 2 x i64> @llvm.aarch64.sve.cdot.lane.nxv2i64(<vscale x 2 x i64> %a,
<vscale x 8 x i16> %b,
<vscale x 8 x i16> %c,
i32 1, i32 270)
ret <vscale x 2 x i64> %out
}
declare <vscale x 4 x i32> @llvm.aarch64.sve.cdot.nxv4i32(<vscale x 4 x i32>, <vscale x 16 x i8>, <vscale x 16 x i8>, i32)
declare <vscale x 2 x i64> @llvm.aarch64.sve.cdot.nxv2i64(<vscale x 2 x i64>, <vscale x 8 x i16>, <vscale x 8 x i16>, i32)
declare <vscale x 4 x i32> @llvm.aarch64.sve.cdot.lane.nxv4i32(<vscale x 4 x i32>, <vscale x 16 x i8>, <vscale x 16 x i8>, i32, i32)
declare <vscale x 2 x i64> @llvm.aarch64.sve.cdot.lane.nxv2i64(<vscale x 2 x i64>, <vscale x 8 x i16>, <vscale x 8 x i16>, i32, i32)
|
