I do not think I'm more experienced in GPU audio processing than Nils is. But since I've tried the same thing I do think I have a clue about this 'biz' - it is just true that things may be much better than what I've experienced myself. If they are better then OK: I just won't be a pioneer, and won't receive any laurels. I'm not an Alexander Macedonian either.Uncle E wrote:That's why you'd be better off releasing your GPU experiments now, as opposed to waiting for any other steps in GPU audio development to be taken.
GPU Delay - A delay plugin running on a graphics card
- KVRAF
- 4030 posts since 7 Sep, 2002
- KVRAF
- 4030 posts since 7 Sep, 2002
Sure thing, I (developer I gave this work to) was not manage to get it working acceptably.Amberience wrote:Or because you're lying and you haven't actually managed to do it.Aleksey Vaneev wrote: There is no reason for me to release my GPU plug-in, because it does not work stable enough, and because I have frozen its further development.
- KVRAF
- 4030 posts since 7 Sep, 2002
-
- KVRist
- 180 posts since 19 Apr, 2004 from Espoo, Finland
The whole world knows Russians are evil.Aleksey Vaneev wrote:I thought the whole world thinks Russians are evil.
- KVRAF
- 20893 posts since 22 Nov, 2000 from Southern California
It's a worthwhile goal, especially because your support requirements will be zero if your plug-ins are so universally compatible. I personally don't see a problem with using DSP or GPU cards but I definitely respect the fact that Voxengo plug-ins have always worked flawlessly for me.Aleksey Vaneev wrote:But my vision is probably a bit more conservative, because I won't think this is success until it works on 99% of computer systems. I mean, I do not have any interest in commercial solution that readily works on 30% of current computer systems only: because, frankly saying, this will be one big problem.
-
- KVRian
- 568 posts since 20 Nov, 2003 from Basel, Switzerland
It's a fundamental rule in the universe that this can *never* happen. Don't be naive, Uncle E!Uncle E wrote:...because your support requirements will be zero
--th
I'm the stereo chancellor
- KVRAF
- 20893 posts since 22 Nov, 2000 from Southern California
Please watch your manners, this comment is disrespectful & completely unproductive.Amberience wrote:Or because you're lying and you haven't actually managed to do it.
- KVRAF
- 6478 posts since 16 Dec, 2002
eh? I thought Nils was German.tahome wrote:So... it looks like the Americans were faster again
Now Americans... everybody knows they're *all* stupid.
I'll get me coat.
- KVRAF
- 4030 posts since 7 Sep, 2002
Just to bring this discussion on right track, I would expect from GPU-driven EQ to be able to deliver 20 stereo channels at 44.1kHz (with 4 bands each) with 3% CPU load (on Athlon 64 X2 4000+) - this takes 10% when computed on CPU.
Otherwise this GPU effort will go 'pffft'.
Correction: this takes 5% when using dual-core (10% with a single core). So, I would expect only 3% load from GPU equalizer.
Otherwise this GPU effort will go 'pffft'.
Correction: this takes 5% when using dual-core (10% with a single core). So, I would expect only 3% load from GPU equalizer.
Last edited by Aleksey Vaneev on Sat Oct 07, 2006 9:50 am, edited 1 time in total.
-
- KVRian
- 1480 posts since 14 Jun, 2003
russians cant be very stupid, that soyuz thing goes off without a hitch and people pay to ride in it.
- KVRAF
- 4030 posts since 7 Sep, 2002
-
- KVRian
- 1480 posts since 14 Jun, 2003
you saw that thing about 80 core cpus in the works?
and ill still have my uad-1 cards!
and ill still have my uad-1 cards!
- KVRAF
- 4030 posts since 7 Sep, 2002
