llvm-mirror/test/CodeGen/X86/widen_cast-5.ll
Sanjay Patel e7d64b577c [X86] add an exedepfix entry for movq == movlps == movlpd
This is a 1-line patch (with a TODO for AVX because that will affect
even more regression tests) that lets us substitute the appropriate
64-bit store for the float/double/int domains.

It's not clear to me exactly what the difference is between the 0xD6 (MOVPQI2QImr) and 
0x7E (MOVSDto64mr) opcodes, but this is apparently the right choice.

Differential Revision: http://reviews.llvm.org/D8691

llvm-svn: 235014
2015-04-15 15:47:51 +00:00

13 lines
329 B
LLVM

; RUN: llc < %s -march=x86 -mattr=+sse4.2 | FileCheck %s
; CHECK: movl
; CHECK: movq
; bitcast a i64 to v2i32
define void @convert(<2 x i32>* %dst.addr, i64 %src) nounwind {
entry:
%conv = bitcast i64 %src to <2 x i32>
%xor = xor <2 x i32> %conv, < i32 255, i32 32767 >
store <2 x i32> %xor, <2 x i32>* %dst.addr
ret void
}