llvm-6502/test/CodeGen/AArch64/arm64-build-vector.ll
Tim Northover 29f94c7201 AArch64/ARM64: move ARM64 into AArch64's place
This commit starts with a "git mv ARM64 AArch64" and continues out
from there, renaming the C++ classes, intrinsics, and other
target-local objects for consistency.

"ARM64" test directories are also moved, and tests that began their
life in ARM64 use an arm64 triple, those from AArch64 use an aarch64
triple. Both should be equivalent though.

This finishes the AArch64 merge, and everyone should feel free to
continue committing as normal now.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@209577 91177308-0d34-0410-b5e6-96231b3b80d8
2014-05-24 12:50:23 +00:00

36 lines
1.3 KiB
LLVM

; RUN: llc < %s -march=arm64 -aarch64-neon-syntax=apple | FileCheck %s
; Check that building up a vector w/ only one non-zero lane initializes
; intelligently.
define void @one_lane(i32* nocapture %out_int, i32 %skip0) nounwind {
; CHECK-LABEL: one_lane:
; CHECK: dup.16b v[[REG:[0-9]+]], wzr
; CHECK-NEXT: ins.b v[[REG]][0], w1
; v and q are aliases, and str is preferred against st.16b when possible
; rdar://11246289
; CHECK: str q[[REG]], [x0]
; CHECK: ret
%conv = trunc i32 %skip0 to i8
%vset_lane = insertelement <16 x i8> <i8 undef, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0>, i8 %conv, i32 0
%tmp = bitcast i32* %out_int to <4 x i32>*
%tmp1 = bitcast <16 x i8> %vset_lane to <4 x i32>
store <4 x i32> %tmp1, <4 x i32>* %tmp, align 16
ret void
}
; Check that building a vector from floats doesn't insert an unnecessary
; copy for lane zero.
define <4 x float> @foo(float %a, float %b, float %c, float %d) nounwind {
; CHECK-LABEL: foo:
; CHECK-NOT: ins.s v0[0], v0[0]
; CHECK: ins.s v0[1], v1[0]
; CHECK: ins.s v0[2], v2[0]
; CHECK: ins.s v0[3], v3[0]
; CHECK: ret
%1 = insertelement <4 x float> undef, float %a, i32 0
%2 = insertelement <4 x float> %1, float %b, i32 1
%3 = insertelement <4 x float> %2, float %c, i32 2
%4 = insertelement <4 x float> %3, float %d, i32 3
ret <4 x float> %4
}