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

Re: Performance issues versus CVS

From: Keven Ring <keven_at_mitre.org>
Date: 2004-04-19 22:35:23 CEST

solo turn wrote:

>>Has anyone seen performance issues such as these?
>for import, and full ceckout, i cannot tell, theses operations are so rare. but:
>with standard operations (svn up, svn ci) we have issues like that if we use a windows based
>client (and 99% here use windows clients).
>since 0.33 using linux as client (i just tried with reiserfs) is very fast (10 times as fast as
>windows), the slowest thing seems to be the locking process. solaris as client is also fast. and
>*nix feels approximately the same speed as cvs, but i did not measure it recently as there is not
>much cvs left ;)
>but i guess you used unix as client?
Yes. Fedora Core 1, full install for the client.

I would agree that import is rare, but a full checkout is less rare....

Still, the times between CVS and an SVN server shouldn't be *THAT* far
off. I can understand that the HTTP would be a bit slower due to the
extra chattiness of it, but still not 10x slower!

Granted for small check-ins, or small repositories, performance is
basically irrelevant/non-comparable. But when comparisons *can* be
made, it would be nice to know why there are huge differences..

I'll check out 1.0.2 over the next couple of days and report back..


To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Mon Apr 19 22:36:42 2004

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

This site is subject to the Apache Privacy Policy and the Apache Public Forum Archive Policy.