Re: Memory Fragmentation - Weekly Report

["Andrei Soare" <andrei soare gmail com>]

<-- Week 9 -->

Hello,

Sorry for my late report, I have been caught up in some health problems, but it's all ok for now.

Last week I have been working on adding the fragmentation level to the graph I already have, with time - lifetime. I have done that successfully, but with poor performance, so now I'm trying to implement a skip list structure, which is the fastest for search operations. This will improve speed a lot.

The results - with the poor performance version - were great, this tool will be very helpful. I played with it a little and noticed that signal handling really causes problems. I shall update my blog with new graphs this week, after I finish debugging my skip lists :)

Oh, and I will be finally using gnuplot, it's the best of all I know, because it allows real-time zoom-in without decrease in clarity (as i mentioned earlier, it redraws the graph each time you zoom in/out).

I also plan to tweak a little the memview.py script, stay tuned.

On Mon, Jul 28, 2008 at 12:25 PM, Vincent Untz <vuntz gnome org> wrote:

Replying late (saw no other replies, but maybe there were some private
oens?) 

It's absolutely not an issue to work on Solaris, and fixing issues there
will work as well as fixing issues on Linux. As long as you're happy,
everything is good :-)

Vincent

Thank you Vincent, yours was the first reply :) I will probably look into dtrace after GSoC, because now other tools have priority.

Regards,
Andrei Soare