[PATCH 2 of 4] obsstore: disable garbage collection during initialisation (issue4456)

Martin von Zweigbergk martinvonz at google.com
Mon Dec 1 06:27:26 UTC 2014


On Sat Nov 29 2014 at 11:47:09 PM Pierre-Yves David <
pierre-yves.david at ens-lyon.org> wrote:

>
>
> On 11/29/2014 10:29 PM, Martin von Zweigbergk wrote:
> >
> >     The provide and
> >     interesting speedup (25%).
> >
> >     On my 58758 markers repo:
> >     before: 0.468247 seconds
> >     after:  0.344362 seconds
> >
> >
> > Do you see similar values from 'time hg ...'? Without trying to
> > understand the patch, it sounds like you could just be moving the GC out
> > of the method that's being timed. I just want to make sure it's an
> > overall gain.
>
> I do see similar gain, the overall time is much more unstable, but I get
> about the same benefit 0.05s
>

0.05s is less than 0.12s and more like 10-15% than 25%, assuming you're
still talking about the same baseline of ~0.46s. It's still a significant
speedup, so it seems worthwhile, but you may want to include the
"end-to-end" times in the description.

Thanks for checking and for spending time on this. It's one of the most
noticeable performance problems I see in day-to-day use.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://selenic.com/pipermail/mercurial-devel/attachments/20141201/43ca87a9/attachment.html>


More information about the Mercurial-devel mailing list