File: llvm.r600.dot4.ll

package info (click to toggle)
llvm-toolchain-11 1%3A11.0.1-2
  • links: PTS, VCS
  • area: main
  • in suites: bullseye
  • size: 995,808 kB
  • sloc: cpp: 4,767,656; ansic: 760,916; asm: 477,436; python: 170,940; objc: 69,804; lisp: 29,914; sh: 23,855; f90: 18,173; pascal: 7,551; perl: 7,471; ml: 5,603; awk: 3,489; makefile: 2,573; xml: 915; cs: 573; fortran: 503; javascript: 452
file content (11 lines) | stat: -rw-r--r-- 566 bytes parent folder | download | duplicates (15) 
1
2
3
4
5
6
7
8
9
10
11
 
; RUN: llc -march=r600 -mcpu=redwood -verify-machineinstrs < %s

declare float @llvm.r600.dot4(<4 x float>, <4 x float>) nounwind readnone

define amdgpu_kernel void @test_dp4(float addrspace(1)* %out, <4 x float> addrspace(1)* %a, <4 x float> addrspace(1)* %b) nounwind {
  %src0 = load <4 x float>, <4 x float> addrspace(1)* %a, align 16
  %src1 = load <4 x float>, <4 x float> addrspace(1)* %b, align 16
  %dp4 = call float @llvm.r600.dot4(<4 x float> %src0, <4 x float> %src1) nounwind readnone
  store float %dp4, float addrspace(1)* %out, align 4
  ret void
}