AMD Radeon HD 2900 XT overclocking results

[LowRun]

Initial samples of R600 do have Inter-Layer Dielectric (ILD) implemented. (http://www.semiconductorglossary.com...erm=dielectric)

in short, it is a tech that enables transistors to work under xtreme conditions, the con of this technique is that it will increase power consumption a lot~!

however, it is also said that the mass production version by TSMC does not feature this technique. so just wondering how is the overclocking headroom for retail R2900XT as compared to these samples sent to media, lol.

We definitely need to wait for retail samples reviews from XS members to get the whole picture, who's gonna bite the bullet first, i will let the dust settle before i jump on the bandwagon. And thanks for the info Victorshen.

[onethreehill]

Quick test on new ATI new GC
http://forums.hardwarezone.com/showthread.php?t=1607936

[Parcher]

3DMark and' insufficient by now it needs to see under the games to parity,' of bases as they involve the two cards.... he/she asks her and' how come you/they don't make to ever see a test under the same game, since do these seem truthful?? You excuse the possible erred translation...

P.(Italy)

[King_Of_Pilots]

3DMark and' insufficient by now it needs to see under the games to parity,' of bases as they involve the two cards.... he/she asks her and' how come you/they don't make to ever see a test under the same game, since do these seem truthful?? You excuse the possible erred translation...

P.(Italy)

Se mi dici quello che volevi dire te lo traduco! anche perche' non ho capito neanche io quello che dicevi!

I told him i'd be pleased to help him translating what he wanna say!

[MakenValborg]

3DMark and' insufficient by now it needs to see under the games to parity,' of bases as they involve the two cards.... he/she asks her and' how come you/they don't make to ever see a test under the same game, since do these seem truthful?? You excuse the possible erred translation...

P.(Italy)

Se mi dici quello che volevi dire te lo traduco! anche perche' non ho capito neanche io quello che dicevi!

I told him i'd be pleased to help him translating what he wanna say!

se non ci fossimo noi italiani

sarebbe propio una noia mortale

[King_Of_Pilots]

se non ci fossimo noi italiani

sarebbe propio una noia mortale

Io dico che siamo coloro che danno allegria al mondO!

[Parcher]

Se mi dici quello che volevi dire te lo traduco! anche perche' non ho capito neanche io quello che dicevi!

I told him i'd be pleased to help him translating what he wanna say!

il 3DMark e' insufficiente dopo la storia dell' AA che perderebbe gia a 3x il 15% delle prestazioni, bisognerebbe vedere sotto lo stesso gioco a parità' di piattaforma e di impostazioni, come si comportano le due schede....
La domanda e' :come mai chi ha la scheda non fa mai vedere questo confronto almeno su un gioco ???

saluti
P.

[eXceeded]

^^ ARGH :O, I don't speak Italian!!!

[King_Of_Pilots]

il 3DMark e' insufficiente dopo la storia dell' AA che perderebbe gia a 3x il 15% delle prestazioni, bisognerebbe vedere sotto lo stesso gioco a parità' di piattaforma e di impostazioni, come si comportano le due schede....
La domanda e' :come mai chi ha la scheda non fa mai vedere questo confronto almeno su un gioco ???

saluti
P.

3d mark is not enough because of the fact that the AA will make a loss of 15% overall performance, it is needed also to bench the same game with same platform and settings, to analyze how the two cards bench...
The question is: why who has this ''uknown'' card never shows this kind of camparizon between GPUs in one or more games?

That's how'd i translate it! Fear not non-italian-speaking users!

[MakenValborg]

^^ ARGH :O, I don't speak Italian!!!

mhhh... it's hard to explain.. because I haven't get it at all...
aniway he was asking why they use only 3dmark for the test.

A game test between 8800 gts and 2900 XT ( whit anti-alias and filters) will be more realistic.

I hope you understand

[onethreehill]

Quick test on new ATI new GC
http://forums.hardwarezone.com/showthread.php?t=1607936

ok dokey. upon request by many.......

compare the 8800GTS 640Mb from Leadtek as comparison.......

http://forums.hardwarezone.com/showp...9&postcount=33

[Radeon]

it's on the shelves here in malaysia
about lets see about usd 500

[Noobie]

3d mark is not enough because of the fact that the AA will make a loss of 15% overall performance, it is needed also to bench the same game with same platform and settings, to analyze how the two cards bench...
The question is: why who has this ''uknown'' card never shows this kind of camparizon between GPUs in one or more games?

That's how'd i translate it! Fear not non-italian-speaking users!

We need to see all AA/AF options benched with different games and screenshots to compare quality! Give us all the data available and we'll make our own article! We just need the data!

[ChaosMinionX]

Anything new?

[Sampsa]

X6800 @ 2,93 GHz, Asus P5W-DH Deluxe and Radeon HD 2900 XT with default cooler:







3DMark06 stable 855/1030 MHz.

Next I'll switch to QX6700 and DFI's RD600 motherbard and DI is coming on Thursday.

[blossa]

It will be interesting to see max result on air cooling. Also, if understand it correctly the GPU gets rather hot? I wonder how DI can handle that...

Thanks for sharing!

[Sampsa]

It will be interesting to see max result on air cooling. Also, if understand it correctly the GPU gets rather hot? I wonder how DI can handle that...

Thanks for sharing!

After DI there will be LN2

[maxxxpct]

first run LN2 no mods

dfi p965 s dark- ocz fleex 1200 -jetway hd 2900xt -x6800

[aMp]

Dammit, I just wanna see what's possible for 24/7 on decent water, including power figures.

[blossa]

After DI there will be LN2

That sounds fun! Let's hope that goes better than it did in the hockey for finland and sweden...

[jas420221]

Soooo now that the NDA is lifted, where is all the information?

[Afterburner]

Dammit, I just wanna see what's possible for 24/7 on decent water, including power figures.

Agreed......

[Sampsa]

Here we go, unfortunately this Thursday was public holiday here in Finland so I'll get dry ice next week:



According to AMD's document R600's recommended maximum operating temperature is 105 Celsius and 125 C is the absolute maximum. Beyond this ASIC will be damaged.



R600 XT GPU TDP is 160 watts and each memory chip two watts = total board TDP 225 watts



Shim is little bit higher than GPU itself so a friend of mine made 1 mm copper cold plate for better contact. Unfortunately shim is glued very tight around the R600 and very hard to remove with razor blade. G80's heat spreader is very nice feature when installing container.



Here is my aluminum solid container installed on Radeon HD 2900 XT. I modified heatsink from Radeon X1800 XT to PWM components.



Graphics card installed to test setup: QX6700, DFI RD600, pair of Corsair XMS6400-C3 and Enermax Galaxy 850 W.

Tests next week

[LowRun]

Seems like you're gonna have fun, looking forward for your results.

[Xion X2]

Sampsa--This may be a rather mundane question...

When using that copper coldplate, do you apply thermal material inbetween both contact areas of your block(s) or just inbetween the GPU/coldplate?

And for those of us who aren't machinists, where could we pick up a 1mm piece of copper like that? Lowe's or Home Depot, possibly? Does anyone know the exact specifications of the shim area?