1
0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-11-24 03:33:20 +01:00
llvm-mirror/test/CodeGen/ARM/vrsqrte.ll
Bob Wilson 6db76aaf10 Add support for ARM's Advanced SIMD (NEON) instruction set.
This is still a work in progress but most of the NEON instruction set
is supported.

llvm-svn: 73919
2009-06-22 23:27:02 +00:00

34 lines
1.2 KiB
LLVM

; RUN: llvm-as < %s | llc -march=arm -mattr=+neon > %t
; RUN: grep {vrsqrte\\.u32} %t | count 2
; RUN: grep {vrsqrte\\.f32} %t | count 2
define <2 x i32> @vrsqrtei32(<2 x i32>* %A) nounwind {
%tmp1 = load <2 x i32>* %A
%tmp2 = call <2 x i32> @llvm.arm.neon.vrsqrte.v2i32(<2 x i32> %tmp1)
ret <2 x i32> %tmp2
}
define <4 x i32> @vrsqrteQi32(<4 x i32>* %A) nounwind {
%tmp1 = load <4 x i32>* %A
%tmp2 = call <4 x i32> @llvm.arm.neon.vrsqrte.v4i32(<4 x i32> %tmp1)
ret <4 x i32> %tmp2
}
define <2 x float> @vrsqrtef32(<2 x float>* %A) nounwind {
%tmp1 = load <2 x float>* %A
%tmp2 = call <2 x float> @llvm.arm.neon.vrsqrtef.v2f32(<2 x float> %tmp1)
ret <2 x float> %tmp2
}
define <4 x float> @vrsqrteQf32(<4 x float>* %A) nounwind {
%tmp1 = load <4 x float>* %A
%tmp2 = call <4 x float> @llvm.arm.neon.vrsqrtef.v4f32(<4 x float> %tmp1)
ret <4 x float> %tmp2
}
declare <2 x i32> @llvm.arm.neon.vrsqrte.v2i32(<2 x i32>) nounwind readnone
declare <4 x i32> @llvm.arm.neon.vrsqrte.v4i32(<4 x i32>) nounwind readnone
declare <2 x float> @llvm.arm.neon.vrsqrtef.v2f32(<2 x float>) nounwind readnone
declare <4 x float> @llvm.arm.neon.vrsqrtef.v4f32(<4 x float>) nounwind readnone