Benchmark: Git 1.6.3.3, Hg 1.3.1, Bzr 1.17

[Jean-Francois Roy]

This data is a little bit concerning. It would seem, without digging  
too much into the methodology, that 2a has no appreciable impact on  
speed, although it seems competiive on size.

I'm ready to accept that bzr is slower than git because of Python, but  
its discrepency versus hg is too large...

On another note Jari, unless you created a repository for bzr, the  
clone (or branch) operation is going to be much more expensive than  
without because it will copy the history data to the new branch. If  
you didn't make one, it would be nice to see numbers with a 2a  
repository.

Jean-François Roy

Sent from my iPhone

On 2009-08-16, at 15:45, Jari Aalto <jari.aalto at cante.net> wrote:

> Jari Aalto <jari.aalto at cante.net> writes:
>
>> John Arbash Meinel <john at arbash-meinel.com> writes:
>>
>>> Jari Aalto wrote:
>>>> I ran *small* *scale* local benchmark on the three DVCS's with
>>>> interesting results. The times should not be takes as absolute,  
>>>> but only
>>>> relative to each other. The test system was Debian/Ext3.
>>
>> Tests were run using no special options (the defaults).
>>
>> I rerun with 'bzr init --2a', but I'm sorry to report that there  
>> was no
>> significant difference. The only thing that was much faster was  
>> *second*
>> add command, which imported 7552 chnages. The repository size  
>> increased
>> %9 (not significant I assume) with --2a.
>>
>> I uploaded the data to: http://www.editgrid.com/user/jaalto/vc-test
>> Here in ascii as well.
>
> I added to editgrid relative differences between Hg vs. Git, Bzr vs.  
> Git
> and Bzr vs. Hg.
>
> According to the "apply kernel path, add, commit, clone" values, it  
> would
> indicate that Bzr (with --2a) is currently rougly 36% behind Hg 1.3.1.
>
> Jari
>
>