AMD/ATI, waiting for the right moment to pounce?

[Nyte]

It seems that PCMagazine has gotten their hands on a "special" version of the Radeon HD 2900XT in a Crossfire setup which nearly DOUBLES performance across some games and benchmarks over NVIDIA's 8800 GTX SLI.

The systems they used were direct from Falcon Northwest and so the question is... what about the rest of us? The folks over at Dailytech provide their insight into this interesting find in an editorial.



The link to the original review is here:

http://www.pcmag.com/article2/0,1895,2146999,00.asp

Source: Dailytech

[Casheti]

Wow... this sounds awesome

I knew ATi wouldn't let us down.

[Grings]

theyre only 1gb gddr4 2900xt's, not special versions, and theyre already available seperately http://www.overclockers.co.uk/showpr...=56&subcat=922

[tofu]

These GDDR4 cards are pretty damn fast then.

Keep in mind, the 8800GTX SLI system had an Intel quad core at 3.73GHz, while the 2900XT GDDR4 had its at 3.61Ghz. This tells me that the OEM version of the cards either have high ass clocks or Falcon NW/ATI are rigging benches (I doubt it though )

Why else would the OEM cards have such an outrageous cooler?

EDIT: Just saw the bench picz, the performance deltas are FKN HUGE >< ATI's prolly got some souped up drivers up their sleeves.

[pbmaster]

I won't lie, ATi has some good stuff. I've said before my first card was an ATi, and I loved it. Not a nVidia fanboy..just bought what was best when I was buying.

[Exceededgoku]

Originally Posted by Casheti

Wow... this sounds awesome

I knew ATi wouldn't let us down.

dont mean to burst anyones bubble but SLI wasn't actually working on the SLI system... That is not to say that that makes it anymore valid that ATI won, 8 month older card has less mature drivers :S. I'd choose the ATI system over the nvidia one anyday!

[tofu]

Originally Posted by Exceededgoku

dont mean to burst anyones bubble but SLI wasn't actually working on the SLI system... That is not to say that that makes it anymore valid that ATI won, 8 month older card has less mature drivers :S. I'd choose the ATI system over the nvidia one anyday!

AWWWW

U mean on both the Falcon NW and the D3LL 720C?

[Nyte]

I don't recall reading anywhere where it says SLI was not working.

[EastCoasthandle]

Here are some comparisons found below:

Source

Someone with a GTX SLI did benchmark it and found similar results here

Also note that in other forums the HD 2900XT is doing much better in Vista then the G80. Benchmarks using Vista here. As you can see the ATI arch is more efficient in running games in Vista then G80 arch is. IMHO, I don't see G80 drivers closing the HD 2900XT delta but any thing is possible.

[hv43082]

If this card does actually outperform 8800GTX, than it's good for ATI. On the note of "AMD/ATI, waiting for the right moment to pounce?", definitely NO. I do hope they outperform the GTX to create competition and push Nvidia to release their next generation card.

[MarcusTaz]

I went with the ATI 2900 XT and sold my 7900GTKO SLI setup... The reality is the driver support for Nvidia is horrible. I had it with them and made the switch and boy i am happy I ever did.. no looking back now benchmarks or not...

[WarEagleAU]

Thats the XTX model though, which, as was expected by us fanbois, put a whoopin on Nvidia. However, I digress, how believable are these numbers. W1zzard, care to take a stab at it?

[Conti027]

it took over 6 months for the right moment?

[Kursah]

That Falcon PC better be some of the best no matter if it was NV Sli or ATI CF, for $8,000 dollars...DAMN! That's close to 10X what I paid for my build! That's just stupid rediculous in my eyes. I know that's off topic, but every time I see an overpriced pre-built PC like that it makes me slightly angry. Sure I know Falcon PC's do a very good job, and they better for the price you pay, but I'd never pay that when I would rather build my own, tweak it the way I want and enjoy it how I want, plus I don't mind building PC's as it's not a hard thing to do!

On the ATI Scores, I think that is due to custom drivers, possibly for future release. I think the scores are just amazing. And WarEagle, I would agree about the XTX deal, but it's clearly posted mulitple times that those are XT's (or it's a large scale typo).

Another Falcon Northwest Mach V can mean only one thing—new tech that's been overclocked within an inch of its life. This time it's the first single graphics cards with 1GB of graphics memory, the ATI Radeon HD 2900XT. Pairing ATI Radeon HD 2900XT graphics cards with the new Mach V's ($7,930, direct) takes Vista gaming performance to new, enviable heights.

Granted for the price of the PC, it should be XTX's. And I think if these numbers hold true to real customers of that system or similar builds, then NV will pull out some new amazing next or second gen DX10 card. The "next" thing is always on the horizon. But amazing scores none-the-less, now if we can find someone that is getting that system or building a something similar...

If this is just XT's in CF and GDDR4 with high clocks, then I'm intrigued to see the XTX's in action, and more-so intrigued to see a Pro variant that may be better OC-able than say the 1950pro line, with at least 256-bit memory, and around 2/3's the features/capacity of an XT variant. Both NV and ATI haven't released a variant in that area and I believe that gap between 8800s/8600's, 2600's/2400's is too large to overlook.

[tkpenalty]

I really do think this was the case.... it was very odd when one of my friends reported the HD2900XT outperforming to a large scale.

Well the HD2900XT was using crappy memory anyway so yeah so that my explain the such dramatic difference in performance.

[TheLostSwede]

Or, it could simply be that Falcon Northwest has Radeon HD 2900 XTX cards....
http://www.fudzilla.com/index.php?op...=1760&Itemid=1

[Kursah]

Very interesting...wonder if that article had bad information/type errors/misrepresentation. I'm sure we'll soon see!

[Xaser04]

Damn nvidias Vista drivers are terrible.

The following is taken from the Geforce 8800 round up review on bit-tech.



This shows a single 8800GTX scoring higher than the SLI setup on the link in the op post despit being on a slower overall system.

This is obviously windows XP but just shows how bad SLI is running (or not I would imagine) under vista.

Whilst I would love to believe ATI have these magic drivers under wraps I think the difference in performance is more likely SLI not working / poor vista drivers rather than a massive performance jump for the HD2900XT.

EDIT: Just to add, the following review also from bit-tech just shows the difference (albeit on older drivers) that vista gives compared to XP for a 8800GTX SLI setup:

http://www.bit-tech.net/hardware/200...epheus_Q80_/11

[Ketxxx]

Them results really highlight how craptacular Vista is for gaming.. just look at the 3dm06 results alone, almost 18,000 in XP vs a measly 12.2k in Vista.. Pfft. Admittedly some of that can be put down to driver immaturity, but almost 6000pts? I doubt it.

[EastCoasthandle]

Originally Posted by Xaser04

Damn nvidias Vista drivers are terrible.

The following is taken from the Geforce 8800 round up review on bit-tech.



This shows a single 8800GTX scoring higher than the SLI setup on the link in the op post despit being on a slower overall system.

This is obviously windows XP but just shows how bad SLI is running (or not I would imagine) under vista.

Whilst I would love to believe ATI have these magic drivers under wraps I think the difference in performance is more likely SLI not working / poor vista drivers rather than a massive performance jump for the HD2900XT.

EDIT: Just to add, the following review also from bit-tech just shows the difference (albeit on older drivers) that vista gives compared to XP for a 8800GTX SLI setup:

http://www.bit-tech.net/hardware/200...epheus_Q80_/11

Where is the link inwhich you obtained that graph from? Your bottom link is pretty old which doesn't take into account more recent drivers (that actually perform similarly).

[Xaser04]

My previous graph is from the 8800 roundup on Bit-tech hardware reviews.

Here: http://www.bit-tech.net/hardware/200...ies_round_up/8
However I have also found the following: (a much newer review of a 8800Ultra that shows 8800GTX sli scores against a HD2900XT CF setup)

Company of Heros

The bottom graph (for some reason I can't get the graph to load as an image here) clearly shows the 8800GTX SLI setup scoring above 90fps with 4xAA and 16xAF at a resolution of 2560*1600. Considerably more than the 37fps per the op post benchmarks. (its also higher than the HD2900XT CF setup)

Prey

Again the bottom graph is the one to look at. This time its 65fps per the Op benchmarks and 88.5fps for the Bit-tech review (the latter also includes 4xAA and 16xAF whereas the Op benchmarks don't state whether any is running)

There are no Synthetic benchmarks run to compare 3dmark06.

This is running on Vista Ultimate with the following setup:

Intel Core 2 Extreme X6800 (operating at 2.93GHz -- 11x266MHz); Asus Striker Extreme motherboard (nForce 680i SLI); 2 x 1GB Corsair XMS2-8500C5 (operating in dual channel at DDR2-800 with 4-4-4-12-1T timings); Seagate Barracuda 7200.9 200GB SATA hard drive; Enermax Galaxy DXX 1000W PSU; Windows Vista Ultimate x86; Nvidia nForce standalone drivers version 15.00 WHQL.


So this system scores going on nearly 3x as high as the one in OP post despite being on a slower system overall (stock X6800 vs overclocked Quad core).

Thats says to me something was wrong with the SLI setup in the OP post benchmarks as the difference in results just isn't right

Also I take back my comment that the nvidia vista drivers suck if the bit-tech review is anything to go by they arn't too bad.


One thing I will note however from the bit-tech review is just how well the HD2900XT scales with CF in comparison to either the 8800GTS and the 800GTX at higher resolutions.

[EastCoasthandle]

Originally Posted by Xaser04

My previous graph is from the 8800 roundup on Bit-tech hardware reviews.

Here: http://www.bit-tech.net/hardware/200...ies_round_up/8
However I have also found the following: (a much newer review of a 8800Ultra that shows 8800GTX sli scores against a HD2900XT CF setup)

Company of Heros

The bottom graph (for some reason I can't get the graph to load as an image here) clearly shows the 8800GTX SLI setup scoring above 90fps with 4xAA and 16xAF at a resolution of 2560*1600. Considerably more than the 37fps per the op post benchmarks. (its also higher than the HD2900XT CF setup)

Prey

Again the bottom graph is the one to look at. This time its 65fps per the Op benchmarks and 88.5fps for the Bit-tech review (the latter also includes 4xAA and 16xAF whereas the Op benchmarks don't state whether any is running)

There are no Synthetic benchmarks run to compare 3dmark06.

This is running on Vista Ultimate with the following setup:

Intel Core 2 Extreme X6800 (operating at 2.93GHz -- 11x266MHz); Asus Striker Extreme motherboard (nForce 680i SLI); 2 x 1GB Corsair XMS2-8500C5 (operating in dual channel at DDR2-800 with 4-4-4-12-1T timings); Seagate Barracuda 7200.9 200GB SATA hard drive; Enermax Galaxy DXX 1000W PSU; Windows Vista Ultimate x86; Nvidia nForce standalone drivers version 15.00 WHQL.


So this system scores going on nearly 3x as high as the one in OP post despite being on a slower system overall (stock X6800 vs overclocked Quad core).

Thats says to me something was wrong with the SLI setup in the OP post benchmarks as the difference in results just isn't right

Also I take back my comment that the nvidia vista drivers suck if the bit-tech review is anything to go by they arn't too bad.


One thing I will note however from the bit-tech review is just how well the HD2900XT scales with CF in comparison to either the 8800GTS and the 800GTX at higher resolutions.

You do realize:
-that the GTX is beating the Ultra in SLI by a wide margin, showing something a miss
-the R600 CF still beats the GTS in the majority of the test using Cat 7.5 (currently we are on Cat 7.6).
-This is the R600 512Meg not the 1024 version.

You can't just say "look, look at this benchmark results of the GTX..." when it's clear that how it should perform (ultra, GTX, GTS...in that order) is not whats being shown (when other reviews show this exact trend. At this point, it's hard to validate the GTX SLI performance results at this time.

[Xaser04]

Originally Posted by EastCoasthandle

You do realize:
-that the GTX is beating the Ultra in SLI by a wide margin, showing something a miss
-the R600 CF still beats the GTS in the majority of the test using Cat 7.5 (currently we are on Cat 7.6).
-This is the R600 512Meg not the 1024 version.

You can't just say "look, look at this benchmark results of the GTX..." when it's clear that how it should perform (ultra, GTX, GTS...in that order) is not whats being shown (when other reviews show this exact trend. At this point, it's hard to validate the GTX SLI performance results at this time.

Eh? I think you completely missed the point of what I posted.

In the bit-tech review there are no Ultra SLI results only single Ultra cards. I only posted the review of the Ultra as it had GTX SLI results in it.

The GTS vs HD2900XT has nothing to do with this thread, the point is about the GTX results in the first OP post benchmarks being off by quite a bit.

[EastCoasthandle]

Originally Posted by Xaser04

Eh? I think you completely missed the point of what I posted.

In the bit-tech review there are no Ultra SLI results only single Ultra cards. I only posted the review of the Ultra as it had GTX SLI results in it.

The GTS vs HD2900XT has nothing to do with this thread, the point is about the GTX results in the first OP post benchmarks being off by quite a bit.

Dude, you are not reading the OP correctly. Please re-read the article. Once you do it will be clear we are discussing the R600 1024meg version, not the 512Meg version.
Here let me help you: Link. If you go to their review it will take you to the link I posted above.

[erocker]

*Opinion* This is just the start. In a couple more months ATi's cards will be king.