I know the usual 'it was originally for old processors and it may not do much here [plus thread safe drivers appear to be rare]' but it can't hurt to try. Furthermore, I noticed ioquake3 has edited the original code for SDL hence there should deviation from original q3 GPL. Editing the #ifndef MAC OS out of the way (without it you may be fooled SMP is attempted just with the SMP #define), win64 gives a black screen when attempting to run r_smp 1.
Trying SMP acceleration... WARNING: You enable r_smp at your own risk! Render thread starting ...succeeded. R_SyncRenderThread took 0 ms R_SyncRenderThread took 0 ms R_SyncRenderThread took 0 ms R_SyncRenderThread took 1 ms etc. Anyone that had better luck? On anything on non-Mac OS. Is it even on Mac OS working currently? Does it give any performance increase (or decrease) btw? Is anyone interested in developing it properly? Any further help would be appreciated. Also, NVIDIA drivers appear to bug out when trying their 'Threaded Optimization' option with severe FPS drops (unless it's off). Anyone that has any idea how can that NVIDIA featur be supported properly? I had send them a message on an NVIDIA dev forum but nobody knew. Is it some kind of NVIDIA trade secret with game manufacturers? Because it sure sounds silly how hard it is to find (dev) information about it. _______________________________________________ ioquake3 mailing list ioquake3@lists.ioquake.org http://lists.ioquake.org/listinfo.cgi/ioquake3-ioquake.org By sending this message I agree to love ioquake3 and libsdl.
