AMD: R7xx Speculation

[shiznit]

From "just playing around" with an HD4850 I'd say, the Aniso-ouput hasn't changed compared to RV670. It's still better than Nvidias wrt track markings in racing games, but more prone to shimmering in high-freq content.

well ever since G80 i see no shimmering at all using High Quality in the driver (G7x was terrible), are you saying ATI's aniso still has it? This might be a deal breaker for me if true.

 

[CarstenS]

Not as terrible as in NV4x/G7x, but more prominent than in the DX10-GF - yes. But you shouldn't take my word for it but take a look at it yourself.

 

[digitalwanderer]

So do we know anything performance-wise about the 4870? :|

 

[Lukfi]

HD 4850 + 20% at least. And there were some Crysis tests with an AMD stamp that put the HD 4870 above the GTX 260: http://forums.vr-zone.com/showthread.php?t=292315
VR-Zone was down today for some time, so a mirror of the picture is here: http://www.extrahardware.cz/files/images/novinky/2008/05kveten/lf/amd_487_cf.png

 

[Razor1]

not real.

 

[AlphaWolf]

not real.

The performance or the graph? And you know this because?

 

[Razor1]

graphs, as res goes up for the hd4870 the % of scaling stays the same, thats just inaccurate.

Given the fact that the HD4x series when equal in vantage to comparing nV cards, but in game performance is just below those cards 10%-15%. Those can't be real.

 

[Jaaanosik]

Yet another pictorial of HD4870, here.

Beefy and powerful!

From the page...

In 25th of this month (some say it will announce in July 8th) the first graphic card with GDDR5 memory module will be announced. The ATI Radeon HD 4870 utilized with 800 shader processors and 512MB GDDR5 memory and 256bit memory interface. When it comes out, the price will be as low as 299USD. Actually it is selling in HK with the price tag 320USD. now let’s see its Autopsy pics.

And, it beats GTX 260, in price and performance. Our review will be out in days, keep tuned.

 

[digitalwanderer]

That sounds loverly.

 

[AlphaWolf]

graphs, as res goes up for the hd4870 the % of scaling stays the same, thats just inaccurate.

Given the fact that the HD4x series when equal in vantage to comparing nV cards, but in game performance is just below those cards 10%-15%. Those can't be real.

So you have no idea. thanks for clearing that up.

 

[Dave Baumann]

It's still better than Nvidias wrt track markings in racing games, but more prone to shimmering in high-freq content.

Can you give some examples?

 

[MfA]

Very sparse/clean power regulation section compared to the 9800-GTX+, 3 caps vs. 13.

 

[Razor1]

So you have no idea. thanks for clearing that up.

really? Well even with the latest drivers the HD series any of them don't get anywhere near the scaling figures of the GF's in Crysis. So hmm guess what they are fake, any way you look at it.

http://www.tweaktown.com/articles/1470/8/page_8_benchmarks_crysis/index.html

any more info ya need?

 

[Mintmaster]

MfA, could you answer something that I asked before?

Since you're pretty knowledgeable on this subject, can you answer something for me?

Aside from instruction bandwidth, does it cost much more to have a SIMD processor that can change instructions every clock (like just about every CPU) instead of every 4?

 

[Kaotik]

really? Well even with the latest drivers the HD series any of them don't get anywhere near the scaling figures of the GF's in Crysis. So hmm guess what they are fake, any way you look at it.

http://www.tweaktown.com/articles/1470/8/page_8_benchmarks_crysis/index.html

any more info ya need?

Remember that there's already in public 3 different drivers supporting HD4k-series, and who knows what's available for reviewers etc in addition to those 3 different sets

 

[MfA]

Knowledgeable? Not really ... I just like making guesses. On one hand I'd say it shouldn't really matter, on the other both NVIDIA and ATI make their warps/wavefronts larger than the width of their architecture (even including double pumping on NVIDIA's side). So I'm obviously missing something.

 

[AlphaWolf]

really? Well even with the latest drivers the HD series any of them don't get anywhere near the scaling figures of the GF's in Crysis. So hmm guess what they are fake, any way you look at it.

http://www.tweaktown.com/articles/1470/8/page_8_benchmarks_crysis/index.html

any more info ya need?

and extra memory would never help scaling at higher resolutions right?

 

[Razor1]

its more fillrate bound then memory

ignore the highest res, thats just pointless in those graphs. still looking at 40% vs 55% scaling when memory isn't a hurdle.

 

[Mintmaster]

graphs, as res goes up for the hd4870 the % of scaling stays the same, thats just inaccurate.

Where exactly are the issues with scaling? You're being incredibly vague.

The only common reference points between the graphs are the GTX 280 (Tweaktown's is overclocked) and a loose comparison between 4850 an 4870, single and crossfire. GTX 280 perf makes sense, and SLI is still scaling better than CF.

Given the fact that the HD4x series when equal in vantage to comparing nV cards, but in game performance is just below those cards 10%-15%. Those can't be real.

WTF are you talking about? Which Vantage scores are you referring to? This is only one game: Crysis. There's nothing inexplicable here.

 

[AlexV]

Where exactly are the issues with scaling? You're being incredibly vague.

The only common reference points between the graphs are the GTX 280 (Tweaktown's is overclocked) and a loose comparison between 4850 an 4870, single and crossfire. GTX 280 perf makes sense, and SLI is still scaling better than CF.

WTF are you talking about? Which Vantage scores are you referring to? This is only one game: Crysis. There's nothing inexplicable here.

Leave him alone, he doesn't know, he's just upset over the fact that the RV770 isn't lackluster.