Derek Smart on ATI Driver issues

[andypski]

... which is the major reason the design behind D3D sux. I can't see why they still in DX8 use this Lock/Unlock approach. It's more painful to code for and provides no advantage over the passive interface used in OpenGL.

It avoids implicit memcpys.

-- Daniel

...but there's no real performance benefit because it is the act of serialising all rendering that destroys performance.

Lock = BAD.

ReadPixel(x,y,z) style interface = somewhat better, but still not great, as anything that requires real-time immediate information about the state of data in the rendering target surface = serialisation disaster.

- Andy.

 

[vogel]

... on hardware that uses linear buffers, which doesn't exist anymore.
With todays hardware in DirectX when updating a surface we get:

I thought we were talking about vertex buffer data which was the first that came to my mind when I read "lock/unlock". Sorry, missed the point I guess

-- Daniel

 

[RaolinDarksbane]

Good god, I actually understand the frequency explanation you guys made. I guess 4 years(going for 5 ) in college so far pays off.

 

[Nagorak]

Good god, I actually understand the frequency explanation you guys made. I guess 4 years(going for 5 ) in college so far pays off.

I'm glad it made sense to somebody. I thought they were speaking Japanese.

 

[Humus]

... on hardware that uses linear buffers, which doesn't exist anymore.
With todays hardware in DirectX when updating a surface we get:

I thought we were talking about vertex buffer data which was the first that came to my mind when I read "lock/unlock". Sorry, missed the point I guess

-- Daniel

Heh, vertex buffers is another beast. Lock/Unlock makes a lot more sense there.

 

[Reverend]

He is the only developer with so much to complain about. Given his history I am surprised ATI bothers with the guy at all.

Er, I think it would be more correct to say "He is the only developer with so much to complain about publicly."

What he has done publicly on several occasions doesn't help his "cause" (get his game to work better OOTB on the widest variety of platforms) IMO.

I don't particularly care though if what he has done is right or wrong since he doesn't have a history of good games nor is he influential in the progress of 3D graphics. I'm sure if JC were to have done what Derek Smart has done (well, JC has done this before, but in a more matter-of-fact and diplomatic tone), all these IHVs would pay attention instead of ignore.

 

[Nagorak]

It kind of sounds like DS's problems are more due to lack of coding ability than anything else. He should spend more time working on his code, instead of mouthing off... Now that his game (BC 3000 A.D.) seems to have more or less a bust (and he wants people to pay for a patch?), I kind of question whether he has much of a future in the gaming industry. I'd be more worried about that if I were him.

 

[KimB]

It kind of sounds like DS's problems are more due to lack of coding ability than anything else. He should spend more time working on his code, instead of mouthing off... Now that his game (BC 3000 A.D.) seems to have more or less a bust (and he wants people to pay for a patch?), I kind of question whether he has much of a future in the gaming industry. I'd be more worried about that if I were him.

I doubt that, Nagorak. His complaints were very specific, and seemed very credible to me. The complaints he has are also along the same lines as those expressed by JC about ATI's drivers a number of months ago.

So, are you asking that he work around bugs in ATI's drivers, rather than request that ATI fix those bugs?

 

[DemoCoder]

Even though this may all be correct, I highly recommend anyone who is interested to run a search on www.dejanews.com or google for "Derek Smart" and for kicks, add in "Phd" as a search term.

Derek isn't actually a reliable commentator on many things. As far as everyone can tell, he lied about having a Phd, BS'ed heavily about AI/neural nets in BC3k, and made really bad gaffs about hardware engineering which he supposedly had experience in.

He may be right about ATI's drivers, but I'd rather hear it from John Carmack or Tim Sweeney.

 

[Nagorak]

It kind of sounds like DS's problems are more due to lack of coding ability than anything else. He should spend more time working on his code, instead of mouthing off... Now that his game (BC 3000 A.D.) seems to have more or less a bust (and he wants people to pay for a patch?), I kind of question whether he has much of a future in the gaming industry. I'd be more worried about that if I were him.

I doubt that, Nagorak. His complaints were very specific, and seemed very credible to me. The complaints he has are also along the same lines as those expressed by JC about ATI's drivers a number of months ago.

So, are you asking that he work around bugs in ATI's drivers, rather than request that ATI fix those bugs?

I'm saying ATI driver problems are the least of his concerns. And, yes, he should work around them... Whatever it takes to get the game working. ATI should fix the bugs, but if they're not going to do so, or not any time soon, then he just has to figure out some way to work things out. I really don't think he can afford not to support an entire segment of people with ATI cards just because their drivers are somewhat broken. It's not as if any driver for any company works 100%...and I wouldn't be surprised if almost every game has a few hacks coded into it.

 

[KimB]

Driver workarounds are a big obstacle to actually fixing the drivers.

No game developer should have to work around any bug in anybody's driver. Unfortunately, reality isn't so kind, and only the more popular developers (such as John Carmack) are allowed the luxury of being able to get bugs fixed in a timely manner.

The #1 reason is simply this: If updated drivers fix a previous bug, they will break every game that worked around that bug.

Oh, and Derek Smart said that he was working around bugs in the drivers. That doesn't mean he's happy about it.