mirror of
https://github.com/c64scene-ar/llvm-6502.git
synced 2024-12-15 04:30:12 +00:00
7c82e6a32a
intrinsics. The second instruction(s) to be handled are the vector versions of count set bits (ctpop). The changes here are to clang so that it generates a target independent vector ctpop when it sees an ARM dependent vector bits set count. The changes in llvm are to match the target independent vector ctpop and in VMCore/AutoUpgrade.cpp to update any existing bc files containing ARM dependent vector pop counts with target-independent ctpops. There are also changes to an existing test case in llvm for ARM vector count instructions and to a test for the bitcode upgrade. <rdar://problem/11892519> There is deliberately no test for the change to clang, as so far as I know, no consensus has been reached regarding how to test neon instructions in clang; q.v. <rdar://problem/8762292> git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@160410 91177308-0d34-0410-b5e6-96231b3b80d8
134 lines
4.0 KiB
LLVM
134 lines
4.0 KiB
LLVM
; RUN: llc < %s -march=arm -mattr=+neon | FileCheck %s
|
|
; NB: this tests vcnt, vclz, and vcls
|
|
|
|
define <8 x i8> @vcnt8(<8 x i8>* %A) nounwind {
|
|
;CHECK: vcnt8:
|
|
;CHECK: vcnt.8 {{d[0-9]+}}, {{d[0-9]+}}
|
|
%tmp1 = load <8 x i8>* %A
|
|
%tmp2 = call <8 x i8> @llvm.ctpop.v8i8(<8 x i8> %tmp1)
|
|
ret <8 x i8> %tmp2
|
|
}
|
|
|
|
define <16 x i8> @vcntQ8(<16 x i8>* %A) nounwind {
|
|
;CHECK: vcntQ8:
|
|
;CHECK: vcnt.8 {{q[0-9]+}}, {{q[0-9]+}}
|
|
%tmp1 = load <16 x i8>* %A
|
|
%tmp2 = call <16 x i8> @llvm.ctpop.v16i8(<16 x i8> %tmp1)
|
|
ret <16 x i8> %tmp2
|
|
}
|
|
|
|
declare <8 x i8> @llvm.ctpop.v8i8(<8 x i8>) nounwind readnone
|
|
declare <16 x i8> @llvm.ctpop.v16i8(<16 x i8>) nounwind readnone
|
|
|
|
define <8 x i8> @vclz8(<8 x i8>* %A) nounwind {
|
|
;CHECK: vclz8:
|
|
;CHECK: vclz.i8 {{d[0-9]+}}, {{d[0-9]+}}
|
|
%tmp1 = load <8 x i8>* %A
|
|
%tmp2 = call <8 x i8> @llvm.ctlz.v8i8(<8 x i8> %tmp1, i1 0)
|
|
ret <8 x i8> %tmp2
|
|
}
|
|
|
|
define <4 x i16> @vclz16(<4 x i16>* %A) nounwind {
|
|
;CHECK: vclz16:
|
|
;CHECK: vclz.i16 {{d[0-9]+}}, {{d[0-9]+}}
|
|
%tmp1 = load <4 x i16>* %A
|
|
%tmp2 = call <4 x i16> @llvm.ctlz.v4i16(<4 x i16> %tmp1, i1 0)
|
|
ret <4 x i16> %tmp2
|
|
}
|
|
|
|
define <2 x i32> @vclz32(<2 x i32>* %A) nounwind {
|
|
;CHECK: vclz32:
|
|
;CHECK: vclz.i32 {{d[0-9]+}}, {{d[0-9]+}}
|
|
%tmp1 = load <2 x i32>* %A
|
|
%tmp2 = call <2 x i32> @llvm.ctlz.v2i32(<2 x i32> %tmp1, i1 0)
|
|
ret <2 x i32> %tmp2
|
|
}
|
|
|
|
define <16 x i8> @vclzQ8(<16 x i8>* %A) nounwind {
|
|
;CHECK: vclzQ8:
|
|
;CHECK: vclz.i8 {{q[0-9]+}}, {{q[0-9]+}}
|
|
%tmp1 = load <16 x i8>* %A
|
|
%tmp2 = call <16 x i8> @llvm.ctlz.v16i8(<16 x i8> %tmp1, i1 0)
|
|
ret <16 x i8> %tmp2
|
|
}
|
|
|
|
define <8 x i16> @vclzQ16(<8 x i16>* %A) nounwind {
|
|
;CHECK: vclzQ16:
|
|
;CHECK: vclz.i16 {{q[0-9]+}}, {{q[0-9]+}}
|
|
%tmp1 = load <8 x i16>* %A
|
|
%tmp2 = call <8 x i16> @llvm.ctlz.v8i16(<8 x i16> %tmp1, i1 0)
|
|
ret <8 x i16> %tmp2
|
|
}
|
|
|
|
define <4 x i32> @vclzQ32(<4 x i32>* %A) nounwind {
|
|
;CHECK: vclzQ32:
|
|
;CHECK: vclz.i32 {{q[0-9]+}}, {{q[0-9]+}}
|
|
%tmp1 = load <4 x i32>* %A
|
|
%tmp2 = call <4 x i32> @llvm.ctlz.v4i32(<4 x i32> %tmp1, i1 0)
|
|
ret <4 x i32> %tmp2
|
|
}
|
|
|
|
declare <8 x i8> @llvm.ctlz.v8i8(<8 x i8>, i1) nounwind readnone
|
|
declare <4 x i16> @llvm.ctlz.v4i16(<4 x i16>, i1) nounwind readnone
|
|
declare <2 x i32> @llvm.ctlz.v2i32(<2 x i32>, i1) nounwind readnone
|
|
|
|
declare <16 x i8> @llvm.ctlz.v16i8(<16 x i8>, i1) nounwind readnone
|
|
declare <8 x i16> @llvm.ctlz.v8i16(<8 x i16>, i1) nounwind readnone
|
|
declare <4 x i32> @llvm.ctlz.v4i32(<4 x i32>, i1) nounwind readnone
|
|
|
|
define <8 x i8> @vclss8(<8 x i8>* %A) nounwind {
|
|
;CHECK: vclss8:
|
|
;CHECK: vcls.s8
|
|
%tmp1 = load <8 x i8>* %A
|
|
%tmp2 = call <8 x i8> @llvm.arm.neon.vcls.v8i8(<8 x i8> %tmp1)
|
|
ret <8 x i8> %tmp2
|
|
}
|
|
|
|
define <4 x i16> @vclss16(<4 x i16>* %A) nounwind {
|
|
;CHECK: vclss16:
|
|
;CHECK: vcls.s16
|
|
%tmp1 = load <4 x i16>* %A
|
|
%tmp2 = call <4 x i16> @llvm.arm.neon.vcls.v4i16(<4 x i16> %tmp1)
|
|
ret <4 x i16> %tmp2
|
|
}
|
|
|
|
define <2 x i32> @vclss32(<2 x i32>* %A) nounwind {
|
|
;CHECK: vclss32:
|
|
;CHECK: vcls.s32
|
|
%tmp1 = load <2 x i32>* %A
|
|
%tmp2 = call <2 x i32> @llvm.arm.neon.vcls.v2i32(<2 x i32> %tmp1)
|
|
ret <2 x i32> %tmp2
|
|
}
|
|
|
|
define <16 x i8> @vclsQs8(<16 x i8>* %A) nounwind {
|
|
;CHECK: vclsQs8:
|
|
;CHECK: vcls.s8
|
|
%tmp1 = load <16 x i8>* %A
|
|
%tmp2 = call <16 x i8> @llvm.arm.neon.vcls.v16i8(<16 x i8> %tmp1)
|
|
ret <16 x i8> %tmp2
|
|
}
|
|
|
|
define <8 x i16> @vclsQs16(<8 x i16>* %A) nounwind {
|
|
;CHECK: vclsQs16:
|
|
;CHECK: vcls.s16
|
|
%tmp1 = load <8 x i16>* %A
|
|
%tmp2 = call <8 x i16> @llvm.arm.neon.vcls.v8i16(<8 x i16> %tmp1)
|
|
ret <8 x i16> %tmp2
|
|
}
|
|
|
|
define <4 x i32> @vclsQs32(<4 x i32>* %A) nounwind {
|
|
;CHECK: vclsQs32:
|
|
;CHECK: vcls.s32
|
|
%tmp1 = load <4 x i32>* %A
|
|
%tmp2 = call <4 x i32> @llvm.arm.neon.vcls.v4i32(<4 x i32> %tmp1)
|
|
ret <4 x i32> %tmp2
|
|
}
|
|
|
|
declare <8 x i8> @llvm.arm.neon.vcls.v8i8(<8 x i8>) nounwind readnone
|
|
declare <4 x i16> @llvm.arm.neon.vcls.v4i16(<4 x i16>) nounwind readnone
|
|
declare <2 x i32> @llvm.arm.neon.vcls.v2i32(<2 x i32>) nounwind readnone
|
|
|
|
declare <16 x i8> @llvm.arm.neon.vcls.v16i8(<16 x i8>) nounwind readnone
|
|
declare <8 x i16> @llvm.arm.neon.vcls.v8i16(<8 x i16>) nounwind readnone
|
|
declare <4 x i32> @llvm.arm.neon.vcls.v4i32(<4 x i32>) nounwind readnone
|