Interview with ATI Catalyst SPM
- Thread starter Tahir2
- Start date
Babel-17
Veteran
"What this basically means is I am responsible for prioritizing and defining ATI's software roadmap in terms of what new features need to be included as well as what priority certain software bugs receive."
If I'm reading that right he appears to be the fellow to contact in order to enquire about if/when the promised feature of ssaa for the R9700 will be delivered.
I'm really impressed by what I'm reading about this card and oh so close to buying it. Should ATI deliver on this feature it would clinch the deal.
http://mirror.ati.com/vortal/r300/educational/main.html
Click on "SMOOTHVISION 2.0" and then click on "2".
If I'm reading that right he appears to be the fellow to contact in order to enquire about if/when the promised feature of ssaa for the R9700 will be delivered.
I'm really impressed by what I'm reading about this card and oh so close to buying it. Should ATI deliver on this feature it would clinch the deal.
http://mirror.ati.com/vortal/r300/educational/main.html
Click on "SMOOTHVISION 2.0" and then click on "2".
misae said:[Zardon] How did you go about improving the drivers?
Emphasis mine.
http://www.driverheaven.net/#article_446
Great interview--thanks for the link! I'd have to agree with your emphasis--to support a new architecture I have been *very* impressed with these drivers right out of the chute. From Terry's comments it's obvious that ATI's priorities are in the right place and that the company intends to remain in the technological lead. His kind of thinking is good for everyone in the industry, IMO.
Babel-17 said:
I've asked two techs, one through my work as a reviewer and one as a private person.
I've asked two PR guys, both through my work and I've also asked the guys haging around here and Rage3D.
I've also sent them a lot of questions for an interview and they said they would send it to the appropriate people, two moths later I got a reply saying that all my questions had their answers within ATis own whitepapers (none of them had, I've double checked all of the questions).
And last but not least I've added it in the Catalyst feedback program (not that many people are likely to get replies to those posts).
The closest I've come yet is sirerics answer at Rage3D that ATi hasn't "promised anything" and he didn't know much more about it.
Personally I take their site as a "promise" and I'm getting tired waiting.
But then again with the limited fillrate on the 9700 I dunno really, I'm thinking GF FX will be a better choice if you want SSAA anyways.
Yup, only twice the fillrate of any competitor's available product
More seriously, if SSAA is bandwidth limited rather than fillrate limited, it potentially will perform better on the 9700 than the FX. On the other hand, it looks like the FX will have some form of SSAA out of the box, while the 9700 has none so far.
antlers4 said:
Yup, only twice the fillrate of any competitor's available product
More seriously, if SSAA is bandwidth limited rather than fillrate limited, it potentially will perform better on the 9700 than the FX. On the other hand, it looks like the FX will have some form of SSAA out of the box, while the 9700 has none so far.
Ti4600 2.5 gpix/5 billion AA samples/10 GB/s
9700 Pro 2.5 gpix/16 billion AA samples/19 GB/s
FX 4 gpix/16 billion AA samples/16 GB/s
(I've rounded off all those numbers, besides I don't remember all of them by the decimal)
or do I have my head up my arse again?
SSAA is both fillrate and bandwidth limited, MSAA is basically just bandwidth limited
so to me it would seem like 9700 Pro has a slight upper hand when using MSAA, depending on how effective their respective compression methods and driver implementetions are
but that FX will surely have the upper hand when using higher levels of SSAA and especially at high res
Ante P said:
About right except you confused pixel and texel fillrate.
Code:
Ti4600 1.2 GPix / 2.4 GTex / 4.8 G AA samples / 10.4 GB/s
9700 Pro 2.6 GPix / 2.6 GTex / 15.6 G AA samples / 19.8 GB/s
GF FX 4.0 GPix / 4.0 GTex / 16.0 G AA samples / 16.0 GB/sHyp-X said:
hmm are you sure about those figures?
runnings 2d benches that test pixel fillrate gives you results rather close to 2000 MP/s rather than just 1200...
I assume those pixels doesn't just magically appear
MP/s is the format of the scores
btw there is a bug with all ATi cards that effects the fillrate scores, personally I think it's a general 2d bug, ATi has only told me that they verified the bug but when I asked if it was general or just in this benchmark they just gave me some cryptic answers
if the 9700 Pro in fact had a much higher fillrate then it would probably kick ass in this benchmark too:
but it doesn't
the scores fluxuate pretty much in this benchmark, sometimes Ti4600 is faster and vice versa
so I'd say it's a draw
and when moving large folders with lots of files around the screen at high res 9700 Pro actually feels a little bit sluggish compared ti GF4Ti
also these scores tell a lot:
Quake 3 four.dm Max Quality (no AA/Aniso)
2048x1536
Radeon 9700 Pro: 98 fps
GeForce 4 Ti 4200: 113 fps
(nVidias ow GF FX scores showed this test running at 180+ fps, ie almost twice as fast as a 9700 Pro)
something is up with the fillrate of ATi cards...
also the fact that there's some benchmarks where I actually get a better fps when running at 112x864-1280x1024 with 6x FSAA than running at 1600x1200 without FSAA tells a lot too
of course I know that compression does a lot here, but not THAT much
misae said:
Well, then, I wish DAoC: SI would stop constantly hard-locking my system. I guess I'm just not playing it the way it's meant to be played.
martrox said:
I'm running DX9/CAT 3s on XP Prof and the game locks up very consistently under certain conditions (such as rainfall). I also lose my AA when I zone (a work friend of mine has never gotten AA to work with his 9700 and SI). Very annoying. Since my paladin is 40.5 level, I've been lurking about in Cornwall/Dartmoor/Lyonesse quite a bit, so I'm damn tempted to uninstall SI for the time being (but I'd lose AA altogether, and the new trees, shadows, mip-mapping, shaded water, and the new interface).
John Reynolds:
The emphasis was mine because I thought it was a slightly controversial point. However right now I have little to no problems with the current set of drivers. Most stable is great however I do anticiapte their are still problems. What is great about ATI's current driver situation is that they are on an aggressive schedule and do seem to be listening to user feedback.
Every driver upgrade adds some performance gains, and fixing stability problems. For that I am impressed with ATI and believe ATI have been devoting resources to the issue of drivers since the release of the Radeon 8500.
It is not perfect but it is getting there.
The emphasis was mine because I thought it was a slightly controversial point. However right now I have little to no problems with the current set of drivers. Most stable is great however I do anticiapte their are still problems. What is great about ATI's current driver situation is that they are on an aggressive schedule and do seem to be listening to user feedback.
Every driver upgrade adds some performance gains, and fixing stability problems. For that I am impressed with ATI and believe ATI have been devoting resources to the issue of drivers since the release of the Radeon 8500.
It is not perfect but it is getting there.
John, yes, I also lose AA when I zone ( which sucks) but I have had no lockups at all. And I'm stil in the game for the next 3 months (have a lvl 44 Bard - they just charged my card!), but have cancelled my subscription. AC2 is pretty incredible, very freeform - you seem to be able to go anywhere(well, almost!). Only "problem" is the lack of people. It's pretty quiet, which, come to think of it, isn't so bad . Of course, the Radeon9700 does have a small problem with the pixel shaders, but it's only noticable when porting...... and the graphics are, well, you just have to see them. Running 1024X768, 4X FSAA, 16X AF everything on (cept land frill distance, set to medium as setting to full cuts speed by 2/3's). BTW, to those that play this game, DO NOT check AF, FSAA or triple buffering within game, as you gain nothing (if set in drivers) and lose a bunch of framerate!
. Of course, the Radeon9700 does have a small problem with the pixel shaders, but it's only noticable when porting...... and the graphics are, well, you just have to see them. Running 1024X768, 4X FSAA, 16X AF everything on (cept land frill distance, set to medium as setting to full cuts speed by 2/3's). BTW, to those that play this game, DO NOT check AF, FSAA or triple buffering within game, as you gain nothing (if set in drivers) and lose a bunch of framerate!misae said:
I realized that. I don't necessarily agree/disagree and I'm not sure how such a statement could ever be proven.
Same here. Just my luck that the game I'm spending most of my time with, by far, is also the one giving me the most grief. Damn negative karma.
Ante P said:
I think it's a general bug. The Radeon 9700 just seems a tiny bit more sluggish in 2D than the GeForce4. It doesn't interfere with anything, but it's more like, "I shouldn't have been able to see that window draw."
Anyway, there's one little thing I'd like to say:
I have a bone to pick with the Radeon 9700's anisotropic filtering implementation. It forces bilinear/trilinear filtering on all surfaces. This means that you will get blurry text in some games, depending on your graphics settings.
Similar threads
- Deleted member 2197
- PC Hardware, Software and Displays