[llvm-commits] [llvm] r51060 - /llvm/trunk/lib/Target/X86/README-SSE.txt

[Chris Lattner]

On May 13, 2008, at 12:05 PM, Evan Cheng wrote:

>> +In sse4 mode, we could use insertps to make both better.
>
> This is not clear. Is insertps fastish (I suppose the load folding
> variant is better than 2 instruction)? Right now the x86 backend is
> very reluctant to make use of any extracting and insertion  
> instructions.

I don't know.  I assume so :), it saves a register if nothing else.

-Chris