View Single Post
Old 26 April 2013, 13:03   #115
move.l #$c0ff33,throat

StingRay's Avatar
Join Date: Dec 2005
Location: Berlin/Joymoney
Posts: 6,184
Originally Posted by meynaf View Post
I measured 1% speed difference on my A1230.
Might all be true but you probably have a very lightweight system, imagine people with a system full of patches/eye candy etc. pp., then that difference would be much higher than 1%. You could say that people should not run too many things in the background when they want to play games and I would agree, however, once I kill the system I do not have to care about that anymore. I do have other problems though (e.g. RTG support is gone) so neither solution is perfect.

Originally Posted by meynaf View Post
Appears it's audio code in the vbl indeed. If my v2 test really doesn't freeze anymore, then it's da stupid audio dma wait code. Why does it work flawlessly on my machine and not on others ?
Do you use system-friendly interrupts (AddIntServer) or do you bang VBR+$6c directly? If it's the latter, how do you acknowledge the interrupt? 040/060 systems need either 2 writes to $dff09c or one more custom chip access after the write to $dff09c to acknowledge the interrupt properly, if you don't do that your interrupt code won't work correctly on fast 040/060 systems.
StingRay is offline  
Page generated in 0.04070 seconds with 11 queries