RPCS3/llvm-mirror
1
0
Fork 0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2025-01-31 20:51:52 +01:00
Code Issues Projects Releases Wiki Activity

[X86][SSE] Use shuffle/sext instead of deprecated (+ auto-upgraded) pmovsxwd intrinsic call

Browse Source 
llvm-svn: 270489
This commit is contained in:
Simon Pilgrim 2016-05-23 21:21:38 +00:00
parent e4d5aef619
commit 3b7fd697f3
1 changed files with 4 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

9
test/CodeGen/X86/movgs.ll
View File

@ -59,9 +59,10 @@ entry:
%0 = load i64, i64 addrspace(256)* %p %0 = load i64, i64 addrspace(256)* %p
%tmp2 = insertelement <2 x i64> zeroinitializer, i64 %0, i32 0 %tmp2 = insertelement <2 x i64> zeroinitializer, i64 %0, i32 0
%1 = bitcast <2 x i64> %tmp2 to <8 x i16> %1 = bitcast <2 x i64> %tmp2 to <8 x i16>
%2 = tail call <4 x i32> @llvm.x86.sse41.pmovsxwd(<8 x i16> %1) nounwind readnone %2 = shufflevector <8 x i16> %1, <8 x i16> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%3 = bitcast <4 x i32> %2 to <2 x i64> %3 = sext <4 x i16> %2 to <4 x i32>
ret <2 x i64> %3 %4 = bitcast <4 x i32> %3 to <2 x i64>
ret <2 x i64> %4
} }
; The two loads here both look identical to selection DAG, except for their ; The two loads here both look identical to selection DAG, except for their
@ -90,5 +91,3 @@ entry:
%tmp4 = add i32 %tmp1, %tmp3 %tmp4 = add i32 %tmp1, %tmp3
ret i32 %tmp4 ret i32 %tmp4
} }
declare <4 x i32> @llvm.x86.sse41.pmovsxwd(<8 x i16>) nounwind readnone