Performance improvements for bzr-2.4 on large trees

[Martitza Mendez]

I forgot to say in this thread what I said in others...I do three complete
experiments and for each step I report the middle value.  Also Windows
memory management gets a lot better if you kill off all essential
processes.  So if I still see s lot of spread in the data I suspect a rogue
process and go hunting for it until I get a tight cluster of results.   As
john points out the really short times are dominated by loading.  So any
variability much larger than that is automatically suspect.
On May 19, 2011 11:10 PM, "Voelker, Bernhard" <
bernhard.voelker at siemens-enterprise.com> wrote:
> Martitza Mendez wrote:
>
>> Platform is Windows XP, 32 bits. As close to zero load as I can get.
>> Just a "DOS" shell and python.
> ...
>> Since I am not seeing as much improvement as John, I speculate that I
might
>> be limited by either memory or filesystem throughput. I will try a
machine
>> manufactured in this decade soon.
>
> XP (and maybe all Windows versions?) is rather bad in memory management
> and IO buffering. Therefore, I usually do e.g. 5 runs one after another
> for the old and the new version of a program, and then disregard the worst
> and the best result for both versions.
>
> Have a nice day,
> Berny
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.ubuntu.com/archives/bazaar/attachments/20110520/3cae426d/attachment.html>