[llvm-commits] [llvm] r162799 - in /llvm/trunk: include/llvm/Analysis/Passes.h include/llvm/Analysis/ProfileDataLoader.h include/llvm/InitializePasses.h include/llvm/LinkAllPasses.h lib/Analysis/Analysis.cpp lib/Analysis/CMakeLists.txt lib/Analysis/ProfileDataLoader.cpp lib/Analysis/ProfileDataLoaderPass.cpp lib/Analysis/ProfileInfo.cpp test/Analysis/Profiling/load-branch-weights-ifs.ll test/Analysis/Profiling/load-branch-weights-loops.ll test/Analysis/Profiling/load-branch-weights-switches.ll

Alastair Murray alastairmurray42 at gmail.com
Thu Aug 30 16:23:49 PDT 2012


Hi Bill,

On 30/08/12 21:24, Bill Wendling wrote:
> The patch looks fine. Thank you for doing it! Let me know if you don't have commit access and I can apply it for you.

I do not have commit access, so if you could please apply it it would be 
appreciated!

> If the namespace issue is messing things up, then you can leave it. It was a minor issue. I'm still not 100% convinced about the weight code. I doubt that you're gaining much by expanding to uint64_t and then back to unsigned. If anything, you should probably just use uint64_t directly in the whole function instead of casting (which, gross):
>
> static uint64_t AddCounts(uint64_t A, uint64_t B) {
>    // ...
> }
>
> Would this address your concerns?

Yes, I could do this.  Reading between the lines, however, I could just 
not saturate on overflow but leave it unhandled?  Is that preferable?

I'll try and get the MemoryBuffer patch written tomorrow.

Regards,
Alastair.



More information about the llvm-commits mailing list