On Tue, Feb 23, 2010 at 09:03:08 -0500, Nathan Gray wrote: > I took out the check for darcs-2 executables and managed to run > darcs-benchmark just fine with a darcs-1 executable. I ran the > benchmark on a large darcs-1 format repo. It took four days.
Cool!
Low priority: It may be useful to extend the new variants framework so
that it supports the distinction between old-fashioned and hashed
repositories. Right now, the two variants known are "Default" (which
should probably be called "Hashed") and "OptimizePristine".
The benefits to this are (a) you get side-by-side tables, which you
just make using your text-editor, so it's not a big deal and (b)
other people can run old-fashioned vs. hashed vs. hashed with optimize
--pristine easily.
> It will be interesting to compare these benchmarks against the hashed
> and darcs-2 format benchmarks (might take a few days to run).
I'd be very eager to see this
> I tried to add a benchmark for 'darcs add', but so far have been
> unsuccessful. I can send my attempt if someone wants to (gently)
> mentor me in haskell.
Sure! Why don't you darcs send a patch?
Thanks for this work! :-)
Eric
PS. Right now, the one idea I want (somebody [*]) to work on is to add
long-term storage of results so that you can display cumulative
benchmarks over the long term. You'd have JSON files like
~/.darcs-benchmark/2010-02-22T14:32
~/.darcs-benchmark/2010-02-22T17:15
~/.darcs-benchmark/2010-02-23T14:32
And if you run darcs-benchmark report, it would output tables
for all the benchmarks it collected. After that, maybe we could use
criterion to do some more principled analysis of the results.
[*] Not necessarily 'else'
--
Eric Kow <http://www.nltg.brighton.ac.uk/home/Eric.Kow>
PGP Key ID: 08AC04F9
signature.asc
Description: Digital signature
_______________________________________________ darcs-users mailing list [email protected] http://lists.osuosl.org/mailman/listinfo/darcs-users
