[PATCH] D15792: [AArch64] Turn off PredictableSelectIsExpensive on the Cortex-A57

[James Molloy via llvm-commits]

Agreed with Renato - for the avoidance of doubt, my rejection is simply
"rejected for now", not "rejected in concept". I'm not convinced about the
numbers that were given and therefore the conclusion from those numbers.

I'd be happy to see a new set of numbers.

James

On Thu, 7 Jan 2016 at 12:14 Renato Golin via llvm-commits <
llvm-commits at lists.llvm.org> wrote:

> rengolin added a comment.
>
> In http://reviews.llvm.org/D15792#321176, @flyingforyou wrote:
>
> > James, I still think this commit might be worth for Cortex-A57. Could
> you give me a second chance, please?
>
>
> The status is "needs revision", not "refused". :)
>
> If you get new numbers that prove this is a good move, you should
> definitely try again.
>
> The hint is to do yourself the analysis that James did. Look at the
> numbers with care, make sure that they're consistent and if they're not,
> investigate why not, and only publicise your findings when you're sure that
> every corner was swept and you have an answer to all issues with the data.
>
> In a nutshell, always distrust your own numbers until you prove yourself
> that they're good.
>
> cheers,
> --renato
>
>
> http://reviews.llvm.org/D15792
>
>
>
> _______________________________________________
> llvm-commits mailing list
> llvm-commits at lists.llvm.org
> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-commits
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20160107/87da0b45/attachment.html>