ASUS 2900XT Benchin begins

[cadaveca]

That's always the case
Don't see why the scores should be removed that are absolutely great

See my sig.


They ARE great, and nice to see, but it's the equivalent of a Nascar team slicing off 300lbs somehow just before the green flag...nobody else has a chance to compete with such, so makes it less important. Use what everyone else does, and win, and that shows your true skill.

Top 10 closed to guys working for OEM's isn't gonna get new high-end users interested...because even money cannot help them compete with non-released products. It's nice to see scores, really it is, but i gotta side with Guatam on this one...




ANyhoo...great job Sampsa. What were the card clocks?

[Kinc]

See my sig.


They ARE great, and nice to see, but it's the equivalent of a Nascar team slicing off 300lbs somehow just before the green flag...nobody else has a chance to compete with such, so makes it less important. Use what everyone else does, and win, and that shows your true skill.

Top 10 closed to guys working for OEM's isn't gonna get new high-end users interested...because even money cannot help them compete with non-released products. It's nice to see scores, really it is, but i gotta side with Guatam on this one...

ANyhoo...great job Sampsa. What were the card clocks?

I agree. Equivalent hardware is needed for it to be considerd as anything else than just a preview of what is in store for the future. Allthough it is nice to be able to have the hardware in advance to get as much time with it as ever posible.

[BeardyMan]

See my sig.


They ARE great, and nice to see, but it's the equivalent of a Nascar team slicing off 300lbs somehow just before the green flag...nobody else has a chance to compete with such, so makes it less important. Use what everyone else does, and win, and that shows your true skill.

Top 10 closed to guys working for OEM's isn't gonna get new high-end users interested...because even money cannot help them compete with non-released products. It's nice to see scores, really it is, but i gotta side with Guatam on this one...




ANyhoo...great job Sampsa. What were the card clocks?

My point was rather:
Person a can afford this
person b can afford a little more
person c can't afford a coockie

the orb has never been the same when it comes to use hardware

[Atrin]

Kinc

i want to know why u dont use HD2900XT 1GB?

[SSXeon]

Im really late to the party, awsome clocks Kinc, very very nice.

[BenchZowner]

Kinc

i want to know why u dont use HD2900XT 1GB?

Because it practically ends up slower than the 512MB versions due to the VRAM.

[Atrin]

Hmm oks thx

[bl1zZ]

I wont be posting anymore for a while now but I see that I forgot to post the Aquamark score here. Done with 2900XT Crossfire.

Any Systemspecs?
CPU/GPU clocks?
Memory clocks?

Best Regards
Dennis

[aNoN_]

I dont get it, how can two 2900xt in crossfire get better score than two 8800Ultras in SLI? I thought the ultras was faster then the 2900xt's

[massman]

It's all about the CPU

[NH|Delph1]

It's all about the CPU

And the board

//Andreas

[Gautam]

SLI's implementation has got some massive CPU overhead which really shows up in AM3. Surprisingly enough you'll find that a single 8800 Ultra is competitive with two 2900XT's in CF, but SLI does significantly worse.

[Kinc]

So it is finaly released, the QX9650 I have uploaded my best saved result so far. Memclock is very low just above 500MHz 4-4-4-5. I have seriuos problem with the coldbug above 5300MHz (it gets more sensitive to cold the higher it is clocked). At 5388MHz with a better memory multiplier I managed to get a much higher score. But everytime after gt4 it looks up either when the score is shown or in the switch between 3D to 2D.

28574 @ 5300MHz cpu clock: http://service.futuremark.com/compare?3dm06=3482373

I have a new cpu here now that I will start testing today.

[Sampsa]

Also posted 28k score by me and crotale which was run @ Intel Sweden overclocking event in the beginning of October:





http://service.futuremark.com/compare?3dm06=3313085

Although the CPU was running 5,4 GHz but unfortunately without any serious platform tweaks and poor memory performance on X38 (1:1 3-3-3-3) the score is well below Kinc's 28,5k. Luckily there is alot more juice left in the system so more results coming soon from me

Next I'll try 29k

[Trox]

Great News KinC
Waiting to see 30k i hope u can do it.

[piotres]

Hello



Can anybody write me measures of those 3 marked places on photo ?

I need that info trully .

Best Regards
Peter

[biohead]

420mm² (20mm x 21mm)

so diagonal = ~29mm

[piotres]

Everybody say another info ..You said about core right ?

One person said me 29 x 29 mm,other 30.5 x 30.5 mm and You 21 x 20 mm ..so what's truth ?

You say that diagonal is 29 mm ...that's SURE info ??

Peter

[tim-]

diagonal is 29mm, and if you make te easy sqrt(a^2*b^2)=c and remake it to sqrt(c^2/2)=lenght of side since we know the core is a square with same lenght on all sides.

which means, cores side lenght is 20,5mm, diagonal is 29mm,

Last pot i was milling I did 29*29mm center piece 0,75mm thick (to avoid problems with the frame around core. the frame is a little bigger so thats proberbly the 30,5mm you've heard about.

I have send you a better answer via PM.

answer:
your big ??? = 29mm
your small ??? = ~20,5

[jas420221]

I dont get it, how can two 2900xt in crossfire get better score than two 8800Ultras in SLI? I thought the ultras was faster then the 2900xt's

They are, by far.

It's all about the CPU

Nope.

And the board

//Andreas

Strike 2.

The 2900xt is a benchmarking QUEEN, plain and simple people.