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

Re: Subversion 1.9.0-dev FSFS performance tests

From: Ivan Zhakov <ivan_at_visualsvn.com>
Date: Fri, 20 Jun 2014 11:56:59 +0200

> Can you guys dig a little bit deeper here? Seems the performance
> regression that Ivan is seeing deserves some more investigation. With
> Ivan testing http: (and file:) on Windows, and Stefan testing svn: on
> Linux (with different cache settings) it's hard to see where the
> regression comes from.

There are the following main factors that lead to performance regressions:
* log addressing adds two additional files for each revision file
(that is obviously not good for unpacked repositories),
* the whole design relies on the enormous cache size (comparable to
the size of the repository itself).

> Right now, the numbers might just as well indicate that f7 is fast on
> Linux but slow on Windows. Or that it's only fast with svn: but slow
> with http:. That would be "not good".

The numbers do not depend on operating system so much. Stefan Sperling
has results that shows performance degradation on OpenBSD too. From
what I see,
FSFS7 may be fast when repository is packed and there is an enormous
cache size. It is slow when repository is unpacked (default) and cache
size if adequate (default).

Ivan Zhakov
Received on 2014-06-20 11:57:48 CEST

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.