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

Re: svn commit/checkout performance issues (P5)

From: <kfogel_at_collab.net>
Date: 2004-11-29 21:10:50 CET

"Bogdanov, Serge" <serge.bogdanov@intel.com> writes:
> Both transferred more-or-less the same amount of data. I tried to keep
> both workspaces in parallel for the experiments and did 2 commits after
> an initial import. There were some differences though, however they were
> small, the numbers of new and modified files for both systems were in my
> previous message. But in both svn commits (first time I forgot to add
> few new files) the waiting time was huge, I did not measure it for the
> first time but it was of the same order of magnitude.
>
> One thing about the file set was that most of the files were binary:
> libraries, executables and png files.

The only thing I can think is that Subversion is doing binary diffing
but Accurev is not. IOW, Subversion is optimizing for the WAN case,
and Accurev for the LAN (or even local box) case. Since you're
testing on local box, naturally Accurev is winning :-).

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Received on Mon Nov 29 21:16:40 2004

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