[darcs-users] recent darcs performance progress

Eric Kow kowey at darcs.net
Thu Apr 15 09:13:21 UTC 2010


You may be interested in testing Benedikt's patch index work in
progress.  It's not ready for prime-time (or even application to
HEAD), but it could be useful to see if the patch index allows
you to perform darcs annotate or darcs show contents --match 'hash X'
in a reasonable fashion.

The rest of this mail is just keeping apprised of progress in Darcs 2.4,
on the patch tracker, and for the future.

On Wed, Apr 14, 2010 at 17:23:46 -0600, Zooko Wilcox-O'Hearn wrote:
> For what it is worth, the members of the Tahoe-LAFS project are still  
> pretty dissatisfied with darcs's performance.

> darcs query contents --quiet --match "hash  
> 20080103234853-92b7f-966e01e6a40dbe94209229f459988e9dea37013a.gz"  
> "docs/running.html"

This is why I'm looking forward to Benedikt's patch index optimisation
<http://bugs.darcs.net/patch173> to land in Darcs HEAD, hopefully in
time for Darcs 2.5.  I think that commands such as darcs annotate and
darcs show contents will become more useful as a result.

> zooko at localhost:~$ time darcs get --lazy http://tahoe-lafs.org/source/ 
> tahoe-lafs/trunk
> Finished getting.

I take this as an implicit "vote" for the network performance project
which aims to address exactly this problem.

The first stage of this project will be to provide a tarball of the
pristine cache so that darcs get --lazy pretty much only has one file to
fetch...

> Note that we are still using darcs-2.3.0 because that is what is in  
> Ubuntu Lucid (Long-Term Support) and because what I read about darcs-2.4 
> seemed to indicate that it wasn't actually a performance improvement for 
> the tasks that I cared about and might actually be a performance 
> regression.

I visited http://wiki.darcs.net/Benchmarks and looked at the Tahoe
graphs for machines quasar, apricot, vs2 (those that we have graphs
for).

In all of these machines, the difference between darcs annotate timings
is less than a second (on one of them, it was faster).

Have a look at http://wiki.darcs.net/Releases/2.4 as well.  You may
find some of the new features (eg. the interactive hunk editor,
--skip-conflicts) to be compelling enough for an upgrade.  Needless
to say, we're always happy for people to be able to give us feedback
on the latest Darcs.

Some good news for Darcs 2.5:

Petr has work in orbit <http://bugs.darcs.net/patch156> which will
complete the arc started by his summer of code project.  The
hashed-storage stuff was a big chunk of work and it looks like we still
need this finishing blow to fully benefit from it.

He also has another patch <http://bugs.darcs.net/patch196> (porting
David's work over to HEAD) which fixes an issue scaling with respect to
the number of patches in your history.

-- 
Eric Kow <http://www.nltg.brighton.ac.uk/home/Eric.Kow>
PGP Key ID: 08AC04F9
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 197 bytes
Desc: Digital signature
URL: <http://lists.osuosl.org/pipermail/darcs-users/attachments/20100415/305a8673/attachment.pgp>


More information about the darcs-users mailing list