mirror of
https://github.com/RPCS3/llvm-mirror.git
synced 2024-11-23 11:13:28 +01:00
ab043ff680
Essentially the same as the GEP change in r230786.
A similar migration script can be used to update test cases, though a few more
test case improvements/changes were required this time around: (r229269-r229278)
import fileinput
import sys
import re
pat = re.compile(r"((?:=|:|^)\s*load (?:atomic )?(?:volatile )?(.*?))(| addrspace\(\d+\) *)\*($| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$)")
for line in sys.stdin:
sys.stdout.write(re.sub(pat, r"\1, \2\3*\4", line))
Reviewers: rafael, dexonsmith, grosser
Differential Revision: http://reviews.llvm.org/D7649
llvm-svn: 230794
50 lines
2.1 KiB
LLVM
50 lines
2.1 KiB
LLVM
; RUN: llc < %s -march=x86 -mattr=+sse4.2 | FileCheck %s
|
|
; CHECK: mulps
|
|
; CHECK: addps
|
|
|
|
; widen a v3f32 to vfi32 to do a vector multiple and an add
|
|
|
|
define void @update(<3 x float>* %dst, <3 x float>* %src, i32 %n) nounwind {
|
|
entry:
|
|
%dst.addr = alloca <3 x float>* ; <<3 x float>**> [#uses=2]
|
|
%src.addr = alloca <3 x float>* ; <<3 x float>**> [#uses=2]
|
|
%n.addr = alloca i32 ; <i32*> [#uses=2]
|
|
%v = alloca <3 x float>, align 16 ; <<3 x float>*> [#uses=2]
|
|
%i = alloca i32, align 4 ; <i32*> [#uses=6]
|
|
store <3 x float>* %dst, <3 x float>** %dst.addr
|
|
store <3 x float>* %src, <3 x float>** %src.addr
|
|
store i32 %n, i32* %n.addr
|
|
store <3 x float> < float 1.000000e+00, float 2.000000e+00, float 3.000000e+00 >, <3 x float>* %v
|
|
store i32 0, i32* %i
|
|
br label %forcond
|
|
|
|
forcond: ; preds = %forinc, %entry
|
|
%tmp = load i32, i32* %i ; <i32> [#uses=1]
|
|
%tmp1 = load i32, i32* %n.addr ; <i32> [#uses=1]
|
|
%cmp = icmp slt i32 %tmp, %tmp1 ; <i1> [#uses=1]
|
|
br i1 %cmp, label %forbody, label %afterfor
|
|
|
|
forbody: ; preds = %forcond
|
|
%tmp2 = load i32, i32* %i ; <i32> [#uses=1]
|
|
%tmp3 = load <3 x float>*, <3 x float>** %dst.addr ; <<3 x float>*> [#uses=1]
|
|
%arrayidx = getelementptr <3 x float>, <3 x float>* %tmp3, i32 %tmp2 ; <<3 x float>*> [#uses=1]
|
|
%tmp4 = load i32, i32* %i ; <i32> [#uses=1]
|
|
%tmp5 = load <3 x float>*, <3 x float>** %src.addr ; <<3 x float>*> [#uses=1]
|
|
%arrayidx6 = getelementptr <3 x float>, <3 x float>* %tmp5, i32 %tmp4 ; <<3 x float>*> [#uses=1]
|
|
%tmp7 = load <3 x float>, <3 x float>* %arrayidx6 ; <<3 x float>> [#uses=1]
|
|
%tmp8 = load <3 x float>, <3 x float>* %v ; <<3 x float>> [#uses=1]
|
|
%mul = fmul <3 x float> %tmp7, %tmp8 ; <<3 x float>> [#uses=1]
|
|
%add = fadd <3 x float> %mul, < float 0x409EE02900000000, float 0x409EE02900000000, float 0x409EE02900000000 > ; <<3 x float>> [#uses=1]
|
|
store <3 x float> %add, <3 x float>* %arrayidx
|
|
br label %forinc
|
|
|
|
forinc: ; preds = %forbody
|
|
%tmp9 = load i32, i32* %i ; <i32> [#uses=1]
|
|
%inc = add i32 %tmp9, 1 ; <i32> [#uses=1]
|
|
store i32 %inc, i32* %i
|
|
br label %forcond
|
|
|
|
afterfor: ; preds = %forcond
|
|
ret void
|
|
}
|