File: llvm.r600.dot4.ll

package info (click to toggle)
llvm-toolchain-19 1%3A19.1.7-3~deb12u1
  • links: PTS, VCS
  • area: main
  • in suites: bookworm-proposed-updates
  • size: 1,998,492 kB
  • sloc: cpp: 6,951,680; ansic: 1,486,157; asm: 913,598; python: 232,024; f90: 80,126; objc: 75,281; lisp: 37,276; pascal: 16,990; sh: 10,009; ml: 5,058; perl: 4,724; awk: 3,523; makefile: 3,167; javascript: 2,504; xml: 892; fortran: 664; cs: 573
file content (11 lines) | stat: -rw-r--r-- 526 bytes parent folder | download | duplicates (7)
1
2
3
4
5
6
7
8
9
10
11
; RUN: llc -mtriple=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(ptr addrspace(1) %out, ptr addrspace(1) %a, ptr addrspace(1) %b) nounwind {
  %src0 = load <4 x float>, ptr addrspace(1) %a, align 16
  %src1 = load <4 x float>, ptr addrspace(1) %b, align 16
  %dp4 = call float @llvm.r600.dot4(<4 x float> %src0, <4 x float> %src1) nounwind readnone
  store float %dp4, ptr addrspace(1) %out, align 4
  ret void
}