Re: Entries caching & Performance

Brandon Ehle <behle@pipedreaminteractive.com> writes:
> On my working copy of 21011 files in 3528 directories, roughly 81% of
> the time is spent
> doing either IO or CPU intensive operations reading XML data in
> XML_Parse() & XML_ParseBuffer(). Roughly 76% of this time comes from
> svn_wc_entries_read. Talking to sussman and pilchie on IRC revealed
> that we are currently doing caching of entries reading only on commits
> and status. Is anyone working on implementing cached entries readings
> for updates & checkouts? This could possibly eliminate a large portion
> of the time spent during updates.

Philip Martin has done most of the entries caching work; I'm not sure
whether he plans to add more caching soon, or has other priorities.

I suspect the latest comments in

   http://subversion.tigris.org/issues/show_bug.cgi?id=749

are somewhat out of date now :-).

> The other half of the time spent comes from ne_xml_parse. I am not
> sure how this could be made faster short of replacing expat with
> something faster.

Does "half" here really mean the 5% (i.e., 81% - 76%)? If so, it's
not worth worrying about. The entries caching is the big game.

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Tue Nov 26 00:33:12 2002