AMD's 3GHz K10 to break 30,000 3DMark06 (Inq)

[Metroid]

I'm not putting much faith in the article at all. All I made was a simple statement.

Also most people here don't seem to fully understand how 3DMark06 works. The SM2 and SM3 scores would certainly go up significantly in conjunction with this much additional CPU power potentially.

Well at least you got the message, many people here need to get it too. CPU Power increases a lot SM2 & SM3.

[informal]

This could be a duel socket mobo with 2 quad core K10s doing the benches.

Neg.,it's one quad core CPU with 2900 in Crossfire.

Here are the the estimated perf. of Quad Core K10 versus intel's equivalently clocked Penryn:
http://www.techarp.com/showarticle.aspx?artno=434

John says the Quad-Core Opteron systems had already been submitted to the benchmark organizations for testing. The results are all under NDA and will only be released on launch date, September 10, 2007. Because they are under NDA, he cannot reveal the actual results, but he gave us some interesting indications of how the Barcelona will eventually fare against an equivalent Intel processor.

* 20-30% better performance overall
* 170% better performance in some benchmarks

170% better in some benchmarks.

[Stapler]

SM2= 10581
SM3= 11798
CPU= 21000?

SM2+SM3 = 22379

He said the score was 30031. So to be precise AMD Phenom clocked at 3.0 processor needs a score of 7653 points to beat that and that is not impossible.

It's not straight addition.

[Metroid]

http://mekk.homeip.net/3dmarkcalc/3D...culator1.2.zip

Use the calculator. 3Dmark06 aint calculated like that. They have different weights in the final score.

I do not believe on that calculator thing mainly because of enhancements that Phenom may had. The score he wrote is right, the calculator I distrust.

[CraptacularOne]

I do not believe on that calculator thing mainly because of enhancements that Phenom may had. The score he wrote is right, the calculator I distrust.

OK dude, were trying to be nice to you It doesn't matter what optimizations a system has. That calculator is the way 3Dmark06 is a tallied up. The ratios won't EVER change based on system or CPU, it's universal measurement for 06.

[[XC] gomeler]

Calculator has been proven time and time again to be accurate if it's the same one as the XLS file out in the 3DMark section. In regards to this 30k with a 100MHz OC on the R600's.. K10 is either a ing monster or Inq is pulling this out of their ass. I'm going to go with the 2nd option

[thephenom]

Can I get a bottle of salt with this news?

While K10 might be good, I doubt it's THAT good.

[Metroid]

OK dude, were trying to be nice to you It doesn't matter what optimizations a system has. That calculator is the way 3Dmark06 is a tallied up. the ratios won't change based on system or CPU, it's universal measurement for 06.

Believe what you want to. I distrust it and I am not telling it as general conception, before you reply something read it first.

[K404]

ok Metroid..

Heres the 3D06 formula. This is NON NEGOTIABLE, because it came from Futuremark

3DMark06 score =



Alhernative CPU test score formula:
CPU score= (CPU Test 1 x 1523 + CPU Test 2 x 1026) (within 4 points)

[Metroid]

ok Metroid..

Heres the 3D06 formula. This is NON NEGOTIABLE, because it came from Futuremark

3DMark06 score =
2.5 x [1/ ( 1.7/ GS + 0.3/ CPU Score )] /2

In other words, it's 2.5 x weighted harmonic mean of GS and CPU Score.
"GS" is the game score and depends on what hardware you're running.
For Shader Model 3.0 hardware that is capable of running all the tests, it's equal to 0.5 x (SM2S + HDR SM3S).
In other words, it's the average of the SM 2.0 and SM 3.0 + HDR tests.

What if you have hardware that can't do the floating-point blending required for HDR, or doesn't support Shader Model 3.0 at all?
In that case, the overall "GS" value is .75 x SM2S, or 3/4ths of the shader model 2.0 score.

The formula:
damnit..links dead! up in a sec

I got it thanks, very well explained

[RPGWiZaRD]

I don't doubt K10 will be bad but I don't really think it will be THAT good. If it's this good on the other hand then it clearly is the most impressing CPU ever produced, way more successful than what A64 was to Pentium 4. That's since Core 2 Duo is ALREADY good. Now for this to be true Core 2 Duo has to look like Pentium D vs X2 or something. Somehow I doubt this already good architecture (Core 2) gets beaten easily. Using only simple logic tells me this is very difficult to believe altho I'd like to believe it.

K10 has to be like a rough 100%+ faster clock for clock than K8 for this article have any truth in it, I think it sounds a bit much.

[LaGUNaMAN]

... Can I get a bottle of salt with this news?...

^Hehe.. I'd like the same order he's having. But make mine a bigger bottle.

[K404]

I got it thanks, very well explained

ah-hahahaha.


check my post again for the numbers-not-words explanation.

The written explanation is also lifted word for word from FM, so go hassle them about it, not me

[CraptacularOne]

Believe what you want to. I distrust it and I am not telling it as general conception, before you reply something read it first.

Don't tell me what to read if you don't even understand what you are talking about. The way the 06 score is calculated is concrete (as was already explained to you) and will not change EVER.

[madcho]

http://www.overclocking-masters.com/...R-3d/3d-06.jpg

score 27039
cpu : 5119 Mhz
2*x2900xt gpu 1175 mem 950

sm2: 11273
hdr/sm3: 12088

With the sm2 and hdr/sm3 scores, you need about 15000 for the cpu to get over 30k

[kemo]

Ok, I'm not believing this article either, but for the "impossible" guys I'm just going to come out and say...

Remember when the C2D came out? Remember how difficult it was to even hit 20k in 05 at that time? C2D came out, and people were beating the extreme cooled A64 results on air without even trying.

So is this impossible? No, it's not. Is it true? Possibly, possibly not.

p.s.
Shintai, we cannot fully assume that the gpu scores would be lower. Different platform, and IIRC in CF SM3 is VERY cpu limited and as such scales VERY well with cpu speed.

I think we all remember that C2D took it from this
to this

FX62 @ 3.6GHZ + 7900GTX SLI = 13.3K
C2D @ 4.8GHZ + 7900GTX SLI = 16.3K
Same bencher , same ram same GPU power add up faster performace per clock and 1.2GHZ more CPU power
so to sum 50% more CPU power resulted in approximately 20% gain in SM2 and 10% in SM3 = a total of 22.5% gain in 3D ,, call it with a 50% more CPU power resulted in 25% 3D power ,, provided that GPU pushed to the max


So comparing C2Q to K10 when tunning HDs 2900XT on air where a 4.5GHZ C2Q will go almost 21K and K10 will do 30K ,, that is more than 40% gain in 3D ,, you translate it into CPU power provided that C2Q is clocked 50% higher


i know that you don't believe the article but i insist no way that a K10 + 2900XT CF on air will hit 30K ,,, 30k will not be done with this generation of GPUs

[Sparky]

I want to believe it, but I'm a bit dubious about it. Would be nice if it were true, but then it is the INQ....

Bah, just release the CPUs already! Oh, and somebody donate one to me too since I can't afford anything new right now

[kunatz]

Man how I hope this turns out to be true...but I doubt it...

On the other hand, it could explain a lot. Maybe K10 was asskicking all the time, but AMD could'nt launch because it was simply way to hot. Fudzilla reports today, that 1,9GHz Barcelona has 95W TDP, and the INQ reported recently that TDP issues were the main problem of K10 marchitecture all the time. Yeah and INQ & FUD are problems...

But to be realistic, as already said, AMD would not have kept that from public in their freaking financial situation.

[Shintai]

Or maybe its actually tripple crossfire

[Metroid]

Don't tell me what to read if you don't even understand what you are talking about. The way the 06 score is calculated is concrete (as was already explained to you) and will not change EVER.

Well now I see what happen when someone believes in different facts, insulting people will take nowhere, better you read once more the forum rules before make another wasteful post.

[CraptacularOne]

Well now I see what happen when someone believes in different facts, insulting people will take nowhere, better you read once more the forum rules before make another wasteful post.

Yours in not a fact, the 3Dmark06 calculator is the only fact here.

[Metroid]

Yours in not a fact, the 3Dmark06 calculator is the only fact here.

Do not try to get over it, I am not saying about the software I'm saying about believing.

[w0mbat]

Or maybe its actually tripple crossfire

I think thats the secret.

[Metroid]

Or maybe its actually tripple crossfire

Well maybe that is the truth behind all of this. They were actually doing some testing with 3 Graphic cards. So maybe Theo did not notice it.

[grunge100]

How much would another graphics card add to the score?