[Ktechlab-devel] profiling is adictive.

[Julian Bäume]

On Wednesday 18 February 2009 03:45:56 Alan Grimes wrote:
> Take my advice, don't get started on profiling and optimization unless
> you have several days worth of time to waste. =P
>
> The biggest CPU sucker in ktechlab is the linear algebra engine. The
> reason it gets called so much is that non-linear things such as
> transistors need to iterate in order to converge on their instantaneous
> DC conditions... I found several ways to reduce the number of times
> QMatrix[] is called... GCC doesn't seem to be inlining it, that could
> also be a fairly substantial boost. The biggest gain was to convert a
> copy (double) loop into a memcpy, that was a very substantial gain!
>
> The most obvious way to get even more out of it is to convert to use
> liboil... =P
What about using eigen2? The developers just released an interesting interview 
on the dot.
http://dot.kde.org/2009/02/16/interview-eigen-developers-20-release

Since it is in kdesupport, we don't need any external dependencies here. (We 
technically don't even link against any shared library, since eigen2 are just 
header files.)

> I fixed a few memory leaks too but a substantial number of trivial leaks
> remain, (I added comments in the code to point them out...)
>
>
> I would like to set up a design meeting sometime next week where people
> can argue over what direction the KDE4 port should go... (even though
> kde4 looks like the biggest regression in open source history...)
+1 from me. We just need to get our time-zones in sync ;)
I am GMT+1, Zoltan is GMT+2, I think.

bye then
julian
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 489 bytes
Desc: This is a digitally signed message part.
URL: <http://mail.kde.org/pipermail/ktechlab-devel/attachments/20090218/dea074bb/attachment.sig>