search

xrmx / bootchart

merge of bootchart-collector and pybootchartgui
GNU General Public License v2.0
233 stars 87 forks source link

Improve performance of _iter_parse_timed_blocks #61

Closed abone28 closed 9 years ago

abone28 commented 9 years ago

I did some rewrite of _iter_parse_timed_blocks since #60.

In my tests it is almost as fast as non-iterator version now while using less memory. In some tests it is even faster than version 0.14.7. I think, this is because of less pagefaults, but not sure.

xrmx commented 9 years ago

Hi Andrey, thanks for the patch! I think the cpu vs memory tradeoff of using iterators is positive enough to not bother with performance optimizations that makes the code more complicated. There's probably other parts of pybootchartgui.py that may for cpu / memory optimizations. Btw python3 is quite a bit faster than 2.x.