Re: [lisaem] 1.0.0 RC 1 is now available for download

It seems to run ok on my G4 (1.4Ghz) mini, however I have to shut down skins and change the display refresh rate to 12Hz or 8Hz, please try that and let me know. On a 500Mhz G4 it is even slower, but with those mode settings (skins off, 12Hz or better 8Hz refresh) it's workable.

Switching video modes to raw will give you back some speed, the Anti Alias mode with Grays should be avoided if possible as the gray replacement introduces a slowdown. The default AntiAlias mode will be ok, but is slower than raw.

The biggest killer is the display refresh rate. You want it to be as low as tolerable on OS X machines, especially G4's. On a dual core intel box, even 60Hz will do ok, but 20 is better... On the G4's 60Hz will slow them to a crawl.

I suppose I could programatically detect a PPC machine and force down to 12Hz, but then someone might have a very fast dual CPU G5 with a fast video card, so I'd prefer not to.

The big difference between the environments is in the way OS X displays work vs Linux/Windows. On OS X they are double buffered, when you have skins turned on it will wind up being tripple or quadruple buffered.

If you can figure out how to get the display system to work better under OS X, I would be very grateful. :-) There is one way to do this, but it would mean writing OS X only code. Right now it's all done under wxWidgets. I don't know enough about the OS X API's to make this happen any time soon, and the other problem is that whatever I do there, I'll have to rewrite under wxWidgets. If you want to write a native OS X display system for the emulator, I'll happily include it.

The reason you don't see updates until you move this mouse on win32 is related to this. I don't refresh the screen until I absolutely have to since it is so expensive to do so.

Also, one thing to note, if you have the dualparallel ROM installed, it takes a very long time to get past the power on self test, so removing the ROM setting and disabling the expansion cards would be recommended for now. This is fixable in the future, I'll add a hack for bypassing the test on that ROM. I think setting all the expansion slot cards to "Nothing" is enough.

The Lisa clock itself is still wrong, it runs 2x-3x faster on just about every machine I've tried it on, so that's another bug.

Hope that this helps. Received on 2007-04-05 10:54:13