Bench results: R600 -vs- 8800GTX in 3DMark2006.

[chuckshissle]

Link

R600

Apparently they tested the first R600 sold.

Test platform:

Processor: Intel Core 2 Extreme X6800 2.93GHz
Mainboard: Intel i975X
Memory: DDR2-800 2GB
OS: Windows Vista
Driver: Catalyst 7.1(8.33), ForceWare 97.44

The score of R600XTX 1GB is about 12000 while NVIDIA 8800GTX 10500. R600XTX is higher than 8800GTX about 15%. R600XT reached 10000 but 8800GTS is 9000.

 

[kaotao]

When are they going to post some non-3dmark benchies???

Still lookin good though.

 

[snowysnowcones]

When are they going to post some non-3dmark benchies???

Still lookin good though.

Yeah, if you had an R600, why would you only bench it on 3dmark?

Fake maybe?

 

[prozac26]

When are they going to post some non-3dmark benchies???

Maybe it's another level 505 gimmick. 8O

Hopefully soon though.

 

[TheGreatGrapeApe]

When are they going to post some non-3dmark benchies???

Probably not until launch, and you know what that means. :?

I have a feeling that the R600 will be less impressive in current DX9 games, and that like the GF8600 it excels in 3Dmark, so feed people that info to keep them from buying other products.

I'm sure there will be great benefits in many situations where the R600 beats the GF8800 by a near equal percentage, but I have a feeling that it oscilates from undewhelming to just barely beating the GF8800, so why post a 5% gain in Oblivion or something instead of a 15% gain in 3Dmark.

And they speak to the n00bs interested in Bungholiomark, but also don't give away much information to the competition about what they're up against until they're actually in market.

Also I think that both nV and ATi can easily optimize for 3Dmark as a test platform for early samples, whereas for games they may focus on them one at a time and hope for lots of user feedback once they are launched.

Anywhoo, just some thoughts on my part as to why we are only seeing early an apparently strategically 'leaked' 3Dmarks from both companies.

 

[blade85]

"3DMark06 test results in VISTA." :? :?

dont the cards work like trolls on vista??

 

[TheGreatGrapeApe]

Nah drivers have come a long way for both IHVs, they are running at about 95+% of XP speed. There are still issues and holes (like ATi with Linux), but it's alot closer than it was.

 

[tamalero]

Nah drivers have come a long way for both IHVs, they are running at about 95+% of XP speed. There are still issues and holes (like ATi with Linux), but it's alot closer than it was.

you know Mr.ape, I was wondering if the R600 versions will be compatible in crossfire
sort of using a 80nm retail version , with a R600 65nm retail in crossfire

 

[chuckshissle]

"3DMark06 test results in VISTA." :? :?

dont the cards work like trolls on vista??

LOL! I don't know why they are doing 3DMark06 on Vista 32bit. I guess they are benching it using DX10 as the card is suppose to be used. I'm hoping they can do some real benching with popular games in XP 32bit soon.

 

[TheGreatGrapeApe]

you know Mr.ape, I was wondering if the R600 versions will be compatible in crossfire
sort of using a 80nm retail version , with a R600 65nm retail in crossfire

Good question. but I doubt it'll matter. They would likely be mfr process agnostic. The X1300-X1650Pros changed from 90-80nm and they are interchangeable.

The biggest concern for Xfire is shader architecture, and remember they work with asynchronous speeds too so an XTX will still operae as an XTX not clock itself down to XT speeds. But they need equal shader resources so add an XL and they would both act like a crippled XL.