Some timings using bzr

[David Ingamells]

James Westby wrote:
> On Fri, 2008-07-11 at 08:21 +0200, David Ingamells wrote:
>   
>> A simple copy of the repository (including the .bzr directory) from
>> NFS to a local disk on desk01 takes 3.1 seconds and NFS to NFS takes
>> 84 seconds at a low-load time of day. Clearly there is significant
>> room for improvement in bzr when you compare the closest bzr
>> equivalents to this copy  (36 and 187 seconds). From an admittedly
>> very naive perspective, the branch bzr makes is only a copy of the
>> repos (when branch is used without options).
>>     
>
> The branch command does do more than this. Firstly it only pulls
> the revisions out of the repository that are present in the history
> of the branch, so it must walk the revision graph, and can't just
> do a plain copy. Secondly it does some integrity checking so that,
> for example, disk corruption would not propogate to the new branch.
> This obviously leads to some overhead. I'm not sure whether that
> justifies all of the difference that you saw, someone who has profiled
> the operation would be able to tell you if there was any wasted
> effort there, and so the operation could be sped up.
>
> Thanks,
>
> James
>
>
>   
As other correspondence in this thread has already suggested the problem 
may be to do with the deprecated dirstate-tags format of the repos. I 
shall try the tests again with the newer format and report back.

My wording "Clearly ..." could have been better phrased. What I was 
trying to say was that there is quite a difference between the 
theoretical minimum time and the time bzr actually takes where 
optimisations could help. Maybe these optimisations have already 
happened with the newer formats - I'll see in my re-runs of the tests.

Thanks and also for your other mail about bzr upgrade.