1 ; RUN: llc -mtriple=aarch64-linux-gnu -mattr=+sve2 < %s | FileCheck %s
8 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) {
10 ; CHECK: cdot z0.s, z1.b, z2.b, #0
12 %out = call <vscale x 4 x i32> @llvm.aarch64.sve.cdot.nxv4i32(<vscale x 4 x i32> %a,
13 <vscale x 16 x i8> %b,
14 <vscale x 16 x i8> %c,
16 ret <vscale x 4 x i32> %out
19 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) {
20 ; CHECK-LABEL: cdot_d:
21 ; CHECK: cdot z0.d, z1.h, z2.h, #90
23 %out = call <vscale x 2 x i64> @llvm.aarch64.sve.cdot.nxv2i64(<vscale x 2 x i64> %a,
24 <vscale x 8 x i16> %b,
25 <vscale x 8 x i16> %c,
27 ret <vscale x 2 x i64> %out
34 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) {
35 ; CHECK-LABEL: cdot_s_idx:
36 ; CHECK: cdot z0.s, z1.b, z2.b[0], #180
38 %out = call <vscale x 4 x i32> @llvm.aarch64.sve.cdot.lane.nxv4i32(<vscale x 4 x i32> %a,
39 <vscale x 16 x i8> %b,
40 <vscale x 16 x i8> %c,
42 ret <vscale x 4 x i32> %out
46 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) {
47 ; CHECK-LABEL: cdot_d_idx:
48 ; CHECK: cdot z0.d, z1.h, z2.h[1], #270
50 %out = call <vscale x 2 x i64> @llvm.aarch64.sve.cdot.lane.nxv2i64(<vscale x 2 x i64> %a,
51 <vscale x 8 x i16> %b,
52 <vscale x 8 x i16> %c,
54 ret <vscale x 2 x i64> %out
58 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)
59 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)
60 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)
61 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)