diff options
| author | Evan Cheng <evan.cheng@apple.com> | 2006-02-17 00:04:28 +0000 |
|---|---|---|
| committer | Evan Cheng <evan.cheng@apple.com> | 2006-02-17 00:04:28 +0000 |
| commit | b590d3a72b56e06a2c540aeb494ebfd2e579aaf9 (patch) | |
| tree | 6282ab9ed76ab92475b5c90c15486d4f4cb20dfc /llvm/lib | |
| parent | 7e5496d5fe7f109fa7034f3bb3faf27f44cb0fc4 (diff) | |
| download | bcm5719-llvm-b590d3a72b56e06a2c540aeb494ebfd2e579aaf9.tar.gz bcm5719-llvm-b590d3a72b56e06a2c540aeb494ebfd2e579aaf9.zip | |
Remind ourselves to revisit the "pxor vs. xorps/xorpd to clear XMM registers"
issue. Need to do more experiments.
llvm-svn: 26247
Diffstat (limited to 'llvm/lib')
| -rw-r--r-- | llvm/lib/Target/X86/README.txt | 6 |
1 files changed, 6 insertions, 0 deletions
diff --git a/llvm/lib/Target/X86/README.txt b/llvm/lib/Target/X86/README.txt index b794176fa53..5a2da2d3d26 100644 --- a/llvm/lib/Target/X86/README.txt +++ b/llvm/lib/Target/X86/README.txt @@ -437,3 +437,9 @@ It might be better to generate movw al, 8(%edx) when we can spare a register. It reduces code size. + +//===---------------------------------------------------------------------===// + +It's not clear whether we should use pxor or xorps / xorpd to clear XMM +registers. The choice may depend on subtarget information. We should do some +more experiments on different x86 machines. |
