Anand has the details about r520,rv530,rv515

[KimB]

Yes they do, and I'm not seeing what you're claiming.

Sorry, I guess my own bias towards OpenGL games and experience with the 9700 Pro (instead of the 9800 XT) crept in there. Looks like the reality is that typically, the 6600 GT just barely outperforms the 9800 XT in most Direct3D games. It does, however, wipe the floor with the 9800 XT in OpenGL games at most settings.
link

 

[Ailuros]

Sorry, I guess my own bias towards OpenGL games and experience with the 9700 Pro (instead of the 9800 XT) crept in there. Looks like the reality is that typically, the 6600 GT just barely outperforms the 9800 XT in most Direct3D games. It does, however, wipe the floor with the 9800 XT in OpenGL games at most settings.
link

GeForces perform in general better in OpenGL irrelevant of what GPU used. Reason enough for ATI to seriously revamp it's OGL driver.

 

[Karma Police]

GeForces perform in general better in OpenGL irrelevant of what GPU used. Reason enough for ATI to seriously revamp it's OGL driver.

Wasn't one of the reasons nVidia's cards worked better on OGL was because of it architecture? Not just because of the drivers or SM3, but because of how it was structured?

 

[KimB]

Wasn't one of the reasons nVidia's cards worked better on OGL was because of it architecture? Not just because of the drivers or SM3, but because of how it was structured?

I don't think so, not at the present time. It's certainly not the primary reason for ATI to perform so much worse than nVidia in OpenGL games than it does in Direct3D games.

 

[Ailuros]

Wasn't one of the reasons nVidia's cards worked better on OGL was because of it architecture? Not just because of the drivers or SM3, but because of how it was structured?

Wherever stencil shadows come to play (like in Doom3 for instance) the lack of optimizations of ATI's ROPs for stenciling might be one reason for the performance differences:

The ROPs are responsible for such basic operations as Z checking (to decide whether the pixel should actually be written, if it hadn’t been rejected from an earlier compare) and either writing or blending pixels to the frame buffer. The NV40 pipeline has both a Z ROP, which does the Z writing, and a C ROP. The C ROP is a combined Z and Colour ROP. The use of the C ROP is what achieves NV2A’s, NV3x’s, and now NV4x’s optimised Z / Stencil rendering path such that during non-colour rendering situations the C ROP can be utilised to write a second Z/Stencil value per clock cycle, but will be used for colour writes when value need to be written to the frame buffer.

http://www.beyond3d.com/previews/nvidia/nv40/index.php?p=11

The basic pixel output engines remain unchanged from R300. In normal operation each pixel pipeline can output 1 colour value and 1 Z/Stencil value, but there is no optimised Z/Stencil rendering performance.

http://www.beyond3d.com/reviews/ati/r420_x800/index.php?p=9

But it won't explain case scenarios like this:

http://www.xbitlabs.com/articles/video/display/evga-7800gtx_17.html

 

[Karma Police]

I don't think so, not at the present time. It's certainly not the primary reason for ATI to perform so much worse than nVidia in OpenGL games than it does in Direct3D games.

Well, I certainly hope ATi took the opportunity of a new chip to address it's OGL performance.

One of the things I was reminded of when looking at the H.A. numbers was the possible performance increase over past cards in OGL games (X850XT vs R520 in Doom3 for example).

 

[KimB]

Last I heard, ATI had recounted its position to revamp its OpenGL drivers for superior performance, stating that OpenGL is still too small a portion of the market for them to bother with.

 

[Karma Police]

Well, with MS putting more emphasis on D3D and forcing OGL to run with an additional layer in Vista, I can see ATi doing that. But with alot of games using the Doom3 engine, that would be a bad short term choice.

 

[ANova]

Last I heard, ATI had recounted its position to revamp its OpenGL drivers for superior performance, stating that OpenGL is still too small a portion of the market for them to bother with.

Considering Doom 3 is the only current game that suffers in framerates I can't imagine why ATI would feel this way. Why does this keep coming up? I thought the whole world knew by now that ATI kills in D3D and is killed in OGL.

Now, that doesn't mean ATI shouldn't have good OGL performance as well, in fact it would help them greatly in the professional sector, but the point is the 9800 XT is easily on par with the 6600 GT in the majority of todays games.

 

[Moloch]

Considering Doom 3 is the only current game that suffers in framerates I can't imagine why ATI would feel this way. Why does this keep coming up? I thought the whole world knew by now that ATI kills in D3D and is killed in OGL.

Now, that doesn't mean ATI shouldn't have good OGL performance as well, in fact it would help them greatly in the professional sector, but the point is the 9800 XT is easily on par with the 6600 GT in the majority of todays games.

Riddick...

 

[KimB]

Well, with MS putting more emphasis on D3D and forcing OGL to run with an additional layer in Vista, I can see ATi doing that. But with alot of games using the Doom3 engine, that would be a bad short term choice.

Well, from what I've been reading, MS only runs OGL through a wrapper in Vista if you also want to run the program in the Aeroglass interface. So full-screen games should be fine.

 

[ANova]

Riddick...

Yes, and I realize Quake 4 and Prey will add to that list, but the majority still uses D3D. I certainly agree ATI should improve or fix whatever it is that is causing the large difference but this is getting off topic.

 

[Karma Police]

Well, from what I've been reading, MS only runs OGL through a wrapper in Vista if you also want to run the program in the Aeroglass interface. So full-screen games should be fine.

My bad, I meant wrapper, not layer.

 

[Geo]

http://www.theinquirer.net/?article=26275

Of course, Inq, so take it with some salt.

But the interesting points is he thinks the clocks will go up (so now we'll have to downclock to 600/1400 to test Sander ).

Also, he's not expecting 90nm NV high-end until 2006, which I found the most interesting. I begin to get the sense that a 7800Ultra is not going to be enuf to "take" 1800XT.

 

[mapel110]

http://www.digit-life.com/articles2/video/r520-preview.html
another preview

 

[Geo]

http://www.digit-life.com/articles2/video/r520-preview.html
another preview

Exec summary: "Meh, nothing special. Might win by a little, might lose by a little. A little bit better pixel pipes, heading down the Xenos road but not as far. Too bad they didn't hit their clocks, because they were initially shooting much higher than 600."

 

[wireframe]

Exec summary: "Meh, nothing special. Might win by a little, might lose by a little. A little bit better pixel pipes, heading down the Xenos road but not as far. Too bad they didn't hit their clocks, because they were initially shooting much higher than 600."

I'm not sure how you came to that conclusion. I thought their own conclusion was pretty scathing:

As for now, there are no prospects of the R520 leadership, no revolution. The most vital question is the same: the real price and availability of top ATI cards. Shall we see the 600 MHz R520 on the shelves right after the announcement?

Most of the article paints a fairly gloomy picture and I cannot find one sentence that hints at ATI "might win by a little." The only ray of hope in the entire article seems to be the improved memory interface with faster memory, although I am not sure if they are saying the memory advantage is there because it is clocked higher or due to efficiencies of the new memory system.

 

[Arty]

Exec summary: "Meh, nothing special. Might win by a little, might lose by a little. A little bit better pixel pipes, heading down the Xenos road but not as far. Too bad they didn't hit their clocks, because they were initially shooting much higher than 600."

Which is basically a rehashed Anandtech article.

 

[Jawed]

http://www.digit-life.com/articles2/video/r520-preview.html
another preview

Sigh, that's the worst thing I've ever read on DL - normally it's a very good site, but that is just bonkers speculation as far as I can tell.

Bemused and disappointed

Jawed

 

[Geo]

By combining these two sentences from adjacent para:

But in fact, everything will depend on a given application, thus testers will not have a unanimous opinion.

And

There may be parity or the 600 MHz cards may be slightly outperformed.

You don't like my exec summary, that's fine --they are notoriously non-granular, y'know. My full sentence was "Might win by a little, might lose by a little" which I thot (obviously, since I said it) was a fair summation of the two points quoted above.