DirectX9 Beta3 released
- Thread starter DemoCoder
- Start date
- Status
Thanks for the news. It's running a bit late, however, when we have to go through at least one Release Candidate and still hit Final in November.
Just one quick question: Is the caps to expose extra features under VS 2.0 and PS 2.0 changed? (No, I'm not talking about VS 3.0 and PS 3.0).
Just one quick question: Is the caps to expose extra features under VS 2.0 and PS 2.0 changed? (No, I'm not talking about VS 3.0 and PS 3.0).
I doubt there will be caps changes for extra features, I have heard of current caps being extended to expose things like the full program lengths the chip can support.
Derek Smart [3000AD]
Regular
Yeah, and it still has W-Buffer support. Contrary to what some were saying that MS had removed it. I countered that in fact, they hadn't. 
Johnny Rotten
Regular
Derek Smart [3000AD said:]Yeah, and it still has W-Buffer support. Contrary to what some were saying that MS had removed it. I countered that in fact, they hadn't.
Lock Thread!! LOCK THREAD!!
Derek Smart [3000AD]
Regular
LOL!!!!
nah, I'm gonna be mellow for awhile
nah, I'm gonna be mellow for awhile
R300 supports it.
Bigus Dickus
Regular
DaveBaumann said:
Would that mean that Derek has a possible solution for his game engine + Radeon 9xxx problems on at least one of the 9xxx cards? You'd think the dev relations guys would have let him know about this possible route to take for the 9700...?
Well, perhaps the thing to be said is that r300 has the potential to support it. There's no guarantee that they will enable it in the drivers and DX9 has to be here in the first place anyway.
Derek Smart [3000AD]
Regular
Bigus Dickus said:
FP Z buffer is not supported in ANY of the current ATI drivers. Period. And even under DX9, I can't use it.
Apart from that, how does an FP Z buffer solve my problem?
If you were to release the game for DX 9 and assume a floating point Z buffer, it would take care of the Z buffer precision problems you are having. It would still require a separate path than the W buffer.
Only problem is, how many years before you can release a game and assume that people have a floating point Z buffer?
Only problem is, how many years before you can release a game and assume that people have a floating point Z buffer?
Bigus Dickus
Regular
Then that answers the question of why the dev rel guys didn't suggest it. Thank you.Derek Smart [3000AD said:
Out of curiosity, is that because the drivers don't expose it, or because it doesn't help you problem even if working properly?
Well, I thought perhaps when DX9 was released, then a sepate code path in your engine could use fp Z-buffer support on the 9700 to alleviate z fighting issues (I had understood that fp z gave more effective precision, much like a 32 bit z buffer would... perhaps I understood incorrectly?). Since any 9700 owner would likely have DX9 installed, and the codepath would be for that specific card, it wouldn't seem to be a problem for the engine to use such a method if available. Any other solution will be 9xxx specific as well, correct?
That only leaves the 9000 with a problem (which, granted, might be the bigger problem since it's a mainstream card and will be sold in greater volume).
Or, are you saying that fp Z-buffer, even if supported in hardware and drivers, wouldn't help your problem with respect to the 9700 at least?
- Status
Similar threads
