I hope so!
Hi guys,
I'm a longtime retrogamer, having started with to play with emulators almost ten years ago. Finally I've some time to collect also technical data about the vintage gaming hardware. The internet is a wonderful place to look for such things.
However one thing about the SNES has eluded me so far and that is a detailed description of the GSU / GSU-2 (aka SuperFX) chip.
I haven't been able, for example to find a complete list of its assembly instructions or something more about its inner hardware besides some vague description of its principles of operation.
Considering that this chip has been fully emulated for quite some time, can an yone point me to some good, extensive and reliable docs about the chip?
Thanks
Is this a good place to ask for details about the SuperFX ?
Moderator: ZSNES Mods
-
- ZSNES Developer
- Posts: 6747
- Joined: Tue Dec 28, 2004 6:47 am
-
- Seen it all
- Posts: 2302
- Joined: Mon Jan 03, 2005 5:04 pm
- Location: Germany
- Contact:
vSNES | Delphi 10 BPLs
bsnes launcher with recent files list
bsnes launcher with recent files list
Look at the US patent for it, and in the second book of the official SNES programming manual.
The biggest things it had going for it was a (manual) internal instruction cache and a very ingenious pixel-bitplane cache.
Another unique feature was that the dithering (checkerboard) pattern could be automatically handled by hardware. Basically you'd tell it to draw a pixel and it would automatically pick one or the other color based on where the pixel was. If one was the transparent color, you'd get 50% translucency.
Too bad the dithering wasn't 4x4, that would have really been something: 16 colors * 15 colors * 16 "intensities" = 3840 "shades" instead of 16 * 15 = 240 "shades".
The biggest things it had going for it was a (manual) internal instruction cache and a very ingenious pixel-bitplane cache.
Another unique feature was that the dithering (checkerboard) pattern could be automatically handled by hardware. Basically you'd tell it to draw a pixel and it would automatically pick one or the other color based on where the pixel was. If one was the transparent color, you'd get 50% translucency.
Too bad the dithering wasn't 4x4, that would have really been something: 16 colors * 15 colors * 16 "intensities" = 3840 "shades" instead of 16 * 15 = 240 "shades".