File: arm64-simd-scalar-to-vector.ll

package info (click to toggle)
llvm-toolchain-14 1%3A14.0.6-12
  • links: PTS, VCS
  • area: main
  • in suites: bookworm
  • size: 1,496,180 kB
  • sloc: cpp: 5,593,972; ansic: 986,872; asm: 585,869; python: 184,223; objc: 72,530; lisp: 31,119; f90: 27,793; javascript: 9,780; pascal: 9,762; sh: 9,482; perl: 7,468; ml: 5,432; awk: 3,523; makefile: 2,538; xml: 953; cs: 573; fortran: 567
file content (22 lines) | stat: -rw-r--r-- 982 bytes parent folder | download | duplicates (27) 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
 
; RUN: llc < %s -mtriple=arm64-eabi -aarch64-neon-syntax=apple -mcpu=cyclone | FileCheck %s
; RUN: llc < %s -mtriple=arm64-eabi -aarch64-neon-syntax=apple -O0 -fast-isel -mcpu=cyclone | FileCheck %s --check-prefix=CHECK-FAST

define <16 x i8> @foo(<16 x i8> %a) nounwind optsize readnone ssp {
; CHECK: uaddlv.16b h0, v0
; CHECK: rshrn.8b v0, v0, #4
; CHECK: dup.16b v0, v0[0]
; CHECK: ret

; CHECK-FAST: uaddlv.16b
; CHECK-FAST: rshrn.8b
; CHECK-FAST: dup.16b
  %tmp = tail call i32 @llvm.aarch64.neon.uaddlv.i32.v16i8(<16 x i8> %a) nounwind
  %tmp1 = trunc i32 %tmp to i16
  %tmp2 = insertelement <8 x i16> undef, i16 %tmp1, i32 0
  %tmp3 = tail call <8 x i8> @llvm.aarch64.neon.rshrn.v8i8(<8 x i16> %tmp2, i32 4)
  %tmp4 = shufflevector <8 x i8> %tmp3, <8 x i8> undef, <16 x i32> zeroinitializer
  ret <16 x i8> %tmp4
}

declare <8 x i8> @llvm.aarch64.neon.rshrn.v8i8(<8 x i16>, i32) nounwind readnone
declare i32 @llvm.aarch64.neon.uaddlv.i32.v16i8(<16 x i8>) nounwind readnone