projects
/
oota-llvm.git
/ commitdiff
commit
grep
author
committer
pickaxe
?
search:
re
summary
|
shortlog
|
log
|
commit
| commitdiff |
tree
raw
|
patch
| inline |
side by side
(parent:
399610a
)
Add a perf optzn corresponding to PR1033.
author
Chris Lattner
<sabre@nondot.org>
Tue, 5 Dec 2006 18:25:10 +0000
(18:25 +0000)
committer
Chris Lattner
<sabre@nondot.org>
Tue, 5 Dec 2006 18:25:10 +0000
(18:25 +0000)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@32229
91177308
-0d34-0410-b5e6-
96231b3b80d8
lib/Target/X86/README-SSE.txt
patch
|
blob
|
history
diff --git
a/lib/Target/X86/README-SSE.txt
b/lib/Target/X86/README-SSE.txt
index 2685956dc1aabd06a6112e0194aedfaf66d8686e..b80661a5b79594cdfd2d6ec643decd5080f43e04 100644
(file)
--- a/
lib/Target/X86/README-SSE.txt
+++ b/
lib/Target/X86/README-SSE.txt
@@
-18,6
+18,11
@@
Think about doing i64 math in SSE regs.
//===---------------------------------------------------------------------===//
+Bitcast to<->from SSE registers should use movd/movq instead of going through
+the stack. Testcase here: CodeGen/X86/bitcast.ll
+
+//===---------------------------------------------------------------------===//
+
This testcase should have no SSE instructions in it, and only one load from
a constant pool: