[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

Re: Performance benchmarks-1.5 results show significant regression in 1.6

From: Mark Phippard <markphip_at_gmail.com>
Date: Thu, 31 Mar 2011 12:23:01 -0400

On Thu, Mar 31, 2011 at 11:04 AM, Mark Phippard <markphip_at_gmail.com> wrote:

> So while 1.7 has made nice improvements over 1.6, there is still a
> huge regression.  I would blame it on tree conflicts, but would this
> show up in a checkout (maybe because it shares code with update)?
> Might be something for someone to look at.

From IRC:

<Bert> markphip: Know issue that regression 1.6 vs 1.5: Most of that
is tree conflict processing

Bert, I think it might be a bit dismissive to say this was a known
issue. If you were to say it was known that these commands are slower
in 1.6 I would probably agree with you, but I think these tests show a
real problem. If you look at the results for a WC that is like our
trunk, the performance difference is almost not measurable. If you
look at a WC with a ton of folders, the performance difference is
minor. In this one test with tons of files in a single folder the
increase is dramatic. I doubt that was known and I am just saying it
might point us to something in our algorithms that needs to be looked
at. That was the point of these tests.

-- 
Thanks
Mark Phippard
http://markphip.blogspot.com/
Received on 2011-03-31 18:23:33 CEST

This is an archived mail posted to the Subversion Dev mailing list.