Re: ReiserFS reading performance patch

["Vladimir 'φ-coder/phcoder' Serbinenko" <phcoder@gmail.com>]

[-- Attachment #1: Type: text/plain, Size: 1425 bytes --]

On 01.05.2012 19:59, Vladimir 'φ-coder/phcoder' Serbinenko wrote:
> On 27.04.2012 16:05, Albert Zeyer wrote:
>> Hi,
>>
>> A while ago, I improved the grub-mount patches a bit. Mostly:
> You've based your work on abandonded branch. All grub-mount parts were
> merged back in trunk a long time ago. At least half of the work you've
> done was unnecessary with newer version.
>> * Write a simplified compile.py script.
> Custom compile scripts are likely to break often. If there were any
> problem in our compile scripts for your platform it's better to announce
> it here.
>> * Remove nested functions. They don't work on Mac because of stack
>> execution protection.
> As I said, your approach isn't just ugly it results in bugs.
>> * Improve ReiserFS reading performance.
> I wrote reiserfs caching a long time ago but it didn'tmake any sizeable
> difference. As for making offset search a logarithmic, here is a patch
> by few simple modifications of get_item.
Patch contained few unrelated items. Take just reiserfs.c part. With the
patch reading 5G takes around a minute. Few more optimisations are still
possible though like making get_item scan bnode using binary search but
their gain is much smaller and they have no chance of making it into
2.00. While the previously attached patch is more in a gray are and I'm
hesitating.
>


-- 
Regards
Vladimir 'φ-coder/phcoder' Serbinenko



[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 294 bytes --]