Message272432
| Author |
skrah |
| Recipients |
facundobatista, loewis, mark.dickinson, rhettinger, skrah, tim.peters, trex58 |
| Date |
2016年08月11日.09:46:25 |
| SpamBayes Score |
-1.0 |
| Marked as misclassified |
Yes |
| Message-id |
<20160811094615.GA19207@bytereef.org> |
| In-reply-to |
<1470907783.96.0.242477861386.issue27725@psf.upfronthosting.co.za> |
| Content |
On Thu, Aug 11, 2016 at 09:29:44AM +0000, REIX Tony wrote:
> -fno-strict-aliasing -fwrapv for gcc
>
> So, that means that you would get better performance if you applied on Python v2.7 what Python v3.5 did about Py_SIZE(x) .
> However, there are probably other places where the aliasing issue still appears in v2.7 .
I doubt that you'll see any measurable performance difference. The main
inefficiencies in Python aren't that low-level: They are in the interpreter
loop.
> Hummm I'll use -qalias=noansi with XLC and see what happens.
Yes, and also use -fwrapv or similar. Otherwise you might get other issues
that are hard to track down. |
|