Re: [Mixxx-devel] Optimization

2009-05-06 Thread Sean M. Pappalardo
Garth Dahlstrom wrote: > You should see if you can ping John S, he had some SSE optimized EQ code > at one point, the code was lost but he might have some insights about > how to optimize the EQs... Care to share his E-mail address? I can't find it (or too much) doing quick searches of the maili

Re: [Mixxx-devel] Optimization

2009-03-17 Thread Garth Dahlstrom
You should see if you can ping John S, he had some SSE optimized EQ code at one point, the code was lost but he might have some insights about how to optimize the EQs... Guess if you are rolling an optimized Qt build, need to do some static linking on Linux for those libs... Probably just have t

Re: [Mixxx-devel] Optimization

2009-03-17 Thread Sean M. Pappalardo
Adam Davison wrote: > Due to the insane maintainability issues associated with this I would > very much encourage you to make sure there is a significant > improvement in latency associated with this before you dig in :) Oh, of course! That's why I'm asking you guys to list the parts of the code t

Re: [Mixxx-devel] Optimization

2009-03-17 Thread Wes Idell
Hi Sean: I don't post here much (well, this is my first post), though I do lurk in the listserv, and am helping Albert on a small piece of library code. I thought that I would chime in on something of which I'm semi-familiar.. Due to todays processors' complexity, including but not limited to pre

Re: [Mixxx-devel] Optimization

2009-03-17 Thread Adam Davison
Hi Sean, Due to the insane maintainability issues associated with this I would very much encourage you to make sure there is a significant improvement in latency associated with this before you dig in :) Adam 2009/3/16 Sean M. Pappalardo : > Hello. > > For reference, I just ran across this site

[Mixxx-devel] Optimization

2009-03-16 Thread Sean M. Pappalardo
Hello. For reference, I just ran across this site which explains how to optimize C++ code with a focus on assembly: http://www.agner.org/optimize/ Being no stranger to assembly language and having hand-counted instruction cycle timing in the past for performance, I would be interested in helpi