From here
http://www.theinquirer.net/default.aspx?article=41970
and our intrepid news section hounds
http://www.xtremesystems.org/forums/...d.php?t=156932
Printable View
From here
http://www.theinquirer.net/default.aspx?article=41970
and our intrepid news section hounds
http://www.xtremesystems.org/forums/...d.php?t=156932
Andy, I should call a mod in here to give you an infraction for double posting. :p:
(I want to believe it can...but my logical side says, no way with only 2 GPU's)
It's just taken me 5 minutes to work out the Poll option Gautam, I'm surprised I did not zero post rather than double :confused: I guess I'm a bit of a noob in some respects!
:D
Rather an interesting topic tonight though don't you think, thought our experts would like to comment? After all we do have the worlds most knowledgeable 3d bench people posting here.....
Regards
Andy
lol andy hypothesising and all
i think AMD CPUs handling decent negative temps will be key :)
fingers crossed they operate a bit better and can manage at least dryice temps :) :)
500mhz = 7k gain in score :D Crazy if true.
rlyQuote:
Originally Posted by k|ngp|n
birdies starting to talk it up big time :shocked:
Oh Theo what have you started :D
:lol: good one
Thats what is stated in other post. Bet that was 05 and not 06 :DQuote:
Originally Posted by dinos22
Cant say till see real screenshot really
hahah i just saw that other threadQuote:
Originally Posted by k|ngp|n
i hear that September will be a VERY interesting month :D ;):);) :D
It's 06...but it might be triple CrossFire???
let the rumor mill keep spinnin its sweet silk. :D
Who knows but those are some huge numbers if 2006. Rumors are more fun than reality anyways :DQuote:
Originally Posted by Gautam
Dual quads and Trifire. That would make it more belivable for sure :)
From my perspective, being rather old, it is a very rare thing indeed for a none expert 3dbencher with new equipment to be able to go in and snatch the Orb away from the experts using old generation stuff. I cannot think of any time at all. True, eventually the front page of the ORB will fill up with say K7, then Northwood, then K8, then the odd Dothan, then Conroe etc but it always in the hands of the Kingpin, Kinc, Sampsa, Shamino, Hipro, Macci, Oppainter, FUGGER, MickeyMouse's of this world, not Joe bloggs.
So if this is true then 1) I am a bit disappointed but 2) Imagine what can be done in the future from the aforementioned ! :eek:
Mind you I still think it is complete bollox if you pin me down on the matter :p:
Regards
Andy
i think not likely either, but somehow i hope that
1. double pci-e speed
2. double the number of pci-e lanes
3. double the number of cpu's
4. double the number of vidcards
would be double my Core2 system @ 3ghz.
Then, i think it's possible.
But, sounds like '05, if the first three doubles listed above are not met. R600 is definately cpu-limited.
Now, if that was 8800's, i think noone would have the same qualms...
I am going to go out on a limb and vote "yes." I do personally believe that 30k has been done with some AMD setup. Time will tell if I have to bite my words. The fact that Theo said 30k has been broken is enough for me to believe that 30k has been broken.
Maybe it's with 3 cards, maybe its with 4 cards, maybe its with multiple CPU's. Either way I think its been done, so invariably we are going to be seeing very huge scores this year.
I'm saying this totally knowing I'm probably going to look like an utter fool within 2 months. (If I don't already)
Theo specifies the setup concisely so there should be no confusion
if he was lying we will be onto him and he will really shoot himself in the foot if those results cannot be redone when these CPUs are out
ok so KP it was not just 500MHz CPU clocks but rather overclocked cards PLUS possible new driver....as he doesn't say what he used on first run....i assume the drive is the sameQuote:
The particular processor was none other than a single socket Barcelona or Agena FX, call it what you will. The reference motherboard containing RD790 chipset packed two HD 2900XT cards, and the memory installed was Corsair’s Dominator PC2-9136C5D, or the same ones we have been using ever since they came out. There was a Raptor hard drive, and that was about that. OCZ will like the fact that PP&C Quaddie CrossFire PSU was installed in the system.
Windows Me II (Vista) Ultimate was installed as the operating system of choice. For some odd reason, the 32-bit version was installed in a system with 4GB of memory, needless to say the system detected 3.24GB, and benchmarketing commenced.
When running at 2.5GHz, the 3DMark06 score ended at 23.768, so we were thrilled to see such a good score coming from two 512MB cards. This showed the clear potential of this four core processor marchitecture, but the helter skelter ride happened after we overclocked the processor to 3.0GHz.
When clocked at 3.0 GHz and equipped with two overclocked HD2900XT cards in CrossFire, Agena FX or single-core Barcelona smashed an index of 30,000 3DMarks 06. Yes, you’ve read it right - the barrier of 30,000 was passed to, barely, 30.031. We know that there are different systems floating around, and we have no doubt that leaks will start appearing after this article goes live. Knowing what was the overclock of Kentsfield processor, Intel Yorkfield has quite a big job to do. One thing is clear, though. Intel needs Nehalem, and it needs it badly.
To be clear with you, dear readers, two HD2900XT cards were overclocked to 830MHz for the core and 900MHz for the memory, but ominous Catalyst 7.7 drivers were used.
Memory wise, it is particularly interesting to see benchmark scores of DDR2-800 on AMD’s Athlon 64 X2 or now this Phenom FX something. When running in dual channel mode, a theoretical bandwidth of 12.8 GB/s can be reached, but that is a pipe dream - of course.
i also assume that RAM was overclocked as well but i don't know what sort of impact that would have on the scores with these AMD CPUs
if Stephen was around he could shed some light on this subject :D
I hope it can. either way I believe it will outperform K8.
ha it had better beat K8 or AMD just spent all this time and money doing nothing.Quote:
Originally Posted by Jimmer411
Though, I suppose you could say that intel initially wasted its time and money with the P4 release.... But hopefully both companies learned from that mistake.
I believe it's possible, but it has to be tri CF and not 2.
I'll believe it when I see it. If you take the details above as true, it's interesting to me this was done on Vista. Vista & K10 w/IMC = big performance jump? That is an interesting possibility and could very well be true (too bad they left out the CPU scores). FM is saying the upcoming 3DMark will only run on Vista so it looks like we'll be leaving the preferred setup behind, just as we did with DX8/2000 for 3DMark 01 ;)
hmmm..i say, SKT 1207 for this system, or AM2+?
if 1207, I'm sold.
I'm not going to say it's impossible but not probable. I just hope these chips don't coldbug at -25C...
Maybe there should be, maybe there shouldn't be. ;)Quote:
Originally Posted by dinos22
this is what rahul sood said in his blog...i don't think he's going to lie in the position he is...
personally i think this setup was with 3 cards...everytime they have demoed the barcelona is running if we can call it like this "triple crossfire"...:)Quote:
"So, AMD decided to unveil Phenom running at 3.0GHz without showing actual benchmarks. What it showed was a game running smoothly with all details enabled, which makes perfect sense to me. And for the record, if you were to benchmark Phenom at 3GHz you would see that it kicks the living crap out of any current AMD or Intel processor—it is a stone cold killer (at 3GHz, now imagine how it would perform if they could squeek some more juice out of it?).
Thank god we have that smiley so no one can tell what anyone means;)Quote:
Originally Posted by Gautam
well trueQuote:
Originally Posted by Gautam
i am just stating a fact he is giving the infomation as if it is dual GFX setup basically
we will wait and see for "reliable" members here to test these out
too bad AMD is too secretive and doesn't have too many ES CPUs around
integrated memory controller -> coldbug :/Quote:
Originally Posted by [XC] gomeler
NOT alwaysQuote:
Originally Posted by overcrash86
we've had 130nm CPUs rememeber
also some 90nm appeared to have no bug to at least -90C
macci already knows the truth, maybe he can say yes/no :)
Maybe soon we'll need 2 cpu and 3 gpu pots all filled with LN2 :eek: :D
I thinks its a fickle dream of a madman !!! I think AMD need to pull something mighty big out of the bag with these chips but I don't think they will preform this well!
I want to see the screen:rolleyes:
i'm sure he told you tooQuote:
Originally Posted by Sampsa
spit it out Sampsa :D j/k
Na-uh, hypothetical example:Quote:
Originally Posted by overcrash86
I have a 3000+ that cold bugs at 300mhz HT @ -100c
However, I can run @ 299MHz HT @ -100c with a 250/200 memory divider.
I've seen a couple CPUs where the memory frequency is greater than the cold bug frequency, which makes me think that it's not the memory controller.
Ok feel free to prove me wrong.
lol the thread question is very different than the pool ones...
Anyway, i dont know if they archieved 30k, and i think 98% of people also
I don't believe that 2 2900s at those speeds could hit 30k sorry.
Well... seems like people at AMD have what they wanted: people talking about them.
K10 should outperform C2D, but I don't know if it will. First of all, should eradicate the coldbug.
k10 should outperform c2d yes.... but.. intel will prolly just pull some stunt and release nehalem faster than originally planned if k10 performs unexpected good....Quote:
Originally Posted by Fixxxer
as for the 30k in 3dmark06....
i dont think it will happen with a k10@3ghz....
it took shamino and kinc + c2d@5,11ghz to get 27k....
that would make it more than a 50% clock per clock performance difference from c2d to k10.... and i dont think that will happen....
Enough crap. Let's show the official spokespeople.
AMD's Quads to outperform Intel's by over 40%
http://www.youtube.com/watch?v=G_n3wvsfq4Y
3GHZ demo is TRIFIRE:
http://www.youtube.com/watch?v=R7EZmYth6TM
EDIT TO FIX FIRST LINK
I dont see any mention in the second link regarding 3DMark'06 or 30k
that make me ask two questionQuote:
Originally Posted by cadaveca
1) drivers good enough for low resolution benching ?
20 how good they are for gaming ?
remember the quad SLI
Drivers are always optimized for benching as of late, it seems. I've got a few 2900XT systems, XP, Vista, 32-bit and 64, and in most scenarios, having a second card merely hampers performance(except in benches). We don't see this in the popular benches as the driver team quite obviously has spent a considerable AMT of time making these apps work very well.Quote:
Originally Posted by kemo6600
Yes, i do, and see above. Part of the problem with quadSLi, IMHO, was the platform it was on, as well as some really intense driver issues. Vista allows this to be overcome, but seemingly each app must be coded for within the driver for multi-gpu rendering to take full advantage of the gpu power. Seems purely the fault of M$, but of course, you need some good programmers for your drivers too...Quote:
Originally Posted by kemo6600
With that in mind, and knowing how much power these cards draw(or how much they aren't sucking up, and how far they really are from the max they CAN draw), I truly feel that there's more to these cards, and the 8800's than meets the eye. I do reemmber something a long time ago about both R600 and g80 liking about 150mhz pci-e, and I'm sure if you talk to the "TOP benchers", they'll confirm that these cards behave as I imply.
Now, I think, really, that if 2900's can reach 30k, than so can 8800's...i look at this situation as a platform presentation, not just cpu, video, or whatever... mere 15% gains in each area(cpu, video, mobo) make for some really signifigant gains overall.
no way...
Not impossible, just very hard to believe (very, very, very, very, very, very, hard).
//Andreas
You missed the latest and more proper one. ;)Quote:
Originally Posted by cadaveca
http://www.techarp.com/showarticle.a...tno=434&pgno=0
I love this part
"The pics are gone with my stolen laptop, though."
:D
It's most likely true.
I remember fanboy also trolling about 8800 and it's gigantic leap in performance..
at the end their buts got nailed but the impressive results...;)
I say no, even if k10 is 40% faster than kentsfield, it would need to be at at least 3ghz just to keep pace with shamino's 27k record with the 5.1ghz kentsfield. For amd to hit 30k, they would have to either make it scale even better, or hit higher clockspeeds, which I do not feel is possible with their soi process, imo they're very lucky to hit 3ghz in the first place. And all of this is assuming k10 is 40% faster clock:clock than kentsfield, that may be false and mean even higher clockspeeds
but theinquirer has never been a complete amd fan site like amdzone, and the fact that no reputable site like vr-zone or someone like coolaler hasn't posted the same info makes it hard to believe. And don't forget, there was tons of hype for the r600 which none of which was true.Quote:
Originally Posted by BeardyMan
And all of that is just logitech, take a look at my previous post and think, is it really possible? Not to mention, when he oc'd the cpu from 2.5ghz to 3ghz, he got a 26% performance increase while the clockspeed increased by only 20%. Unless you can prove to me that k10 scales better than 100% in 3dmark06, I'm calling bs.
perhaps he actually was robbed if this article is true, but last time I checked only shamino and kinc were robbed
http://www.theinquirer.net/?article=41909
IMHO it's a typo, 3DM05 not '06. That sounds more realistic. Anyway you look on it, AgenaFX@3GHz scoring 30k 3DM'05s is a really superb result for graphics cards clocked so low.
P.S.: It means that overclocking software for crossfired ATi Radeon HD2900XT graphics cards exist, it just isn't public yet :D