[darcs-devel] patch index performance impact on add, record, amend, obliterate

[Aditya]

Hi,
 These tests are interesting. I have made a 10,000 patch repo and have
tested the time taken to amend-record the first patch(40sec non-pi to 50sec
pi), but I have not looked at how long it takes cumulatively.
 One thing I noticed is that the patch modified is at the very top of the
repository. This means that: all the patches read are new patches.
 In these cases, there is an easy quadratic to linear optimization for
update, which I have implemented.
I will let you guys know the result soon.

Aditya


On Sun, Feb 24, 2013 at 3:11 PM, Guillaume Hoffmann <guillaumh at gmail.com>wrote:

> Hi Ganesh,
>
> sorry, I only started to look at the topic of performance in darcs
> 2.10 yesterday.
>
> The scripts will be (probably) updated later at:
>
>     darcs get http://hub.darcs.net/gh/bench
> Guillaume
>



-- 
BSRK Aditya
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osuosl.org/pipermail/darcs-devel/attachments/20130224/a8b65070/attachment.html>