[llvm] r195496 - X86: Perform integer comparisons at i32 or larger.
silvas at purdue.edu
Sat Nov 30 20:21:50 PST 2013
On Sat, Nov 30, 2013 at 1:56 PM, Rafael Espíndola <
rafael.espindola at gmail.com> wrote:
> > I would expect conclusive benchmark results on a large corpus of code
> > on a broad spectrum of microarchitectures from both Intel and AMD before
> > making a change like this. Also, the code size impact needs to be
> > quantified; having to break out to a whole new instruction for what would
> > otherwise be an immediate has the potential for insane code-size increase
> > (clang has something like 30,000 cmpb's and it seems like almost all of
> > have immediate memory operands). Also, not being able to fold a load into
> > the comparison will likely increase register pressure.
> I tested this with a .bc from a LTO build of clang. The size of .text
> goes from 35468517 bytes to 35589941 bytes, so 1.0034 times larger,
> which seems a pretty small code growth.
Thanks for looking into it. Your results seem to confirm Chandler's code
-- Sean Silva
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the llvm-commits