I highly doubt that will happen again, that was very embarrasing for ati for one, and for another, most likely either it was a joke by ati or whoever did was instantly fired the moment he/she was caught
Printable View
I highly doubt that will happen again, that was very embarrasing for ati for one, and for another, most likely either it was a joke by ati or whoever did was instantly fired the moment he/she was caught
512MB ATI HD 4870(RV770), 1GB GeForce 9800 GX2: Crysis ve 3DMark 2006 Testleri
http://forum.donanimhaber.com/m_22869046/tm.htm
Fake :confused:
Fake
Looks Fake because the difference is very Big between both. And HD 4000 series is not that monster I think.
The 4870 GDDR5 should get around 12.5K in 3D 06 at stock, at least according to the specs (and not the fake numbers). As for performance, it should be able to equal the 8800 GT 512MB in the games in which the 3870 lagged behind.
We will also see larger performance gains from overclocking due to the greater number of shaders. Should be interesting, but not mindblowing.
Perkam
It is definately FAKE. Look at the cpu speed, 3 GHz. The 06 score is impossible at that cpu speed, no matter how fast the GPU is.
I am very certain that the 4870 will be better then the 8800 GT.
If Im currently getting 11500 in 3D mark 06, I doubt that the 4870 is only going to be 1000 points better.
Performance of a single 4870 is expected to be just below a 3870 X2, at least thats what a lot of people have been saying so far.
Don't you think you are underestimating it a bit perk? How is it that an architecture with virtually double the amount of features (shades, textures) etc, is supposed to become a mild update of its previous gen counterpart?
There is 0 chance that 4870 will be on the level of 8800 GT. It will come out to compete with 9800 GTX and nothing less. I'm quite surprised you expect so little from this card when you used to expect so much from R600 before it came out ;)
If past experience has taught us anything, those that overestimate ATI and underestimate Nvidia know nothing about this industry :)
Check out the performance in games for the 3870 vs the 8800 GT 512MB.
It is on par for some games like UT3 and HL2, but for most it lags behind by an average of 30% or so. The 4870 will close that gap, beat the 9800 GTX in 3d 06 (by 1000 or so points), and overclock better than current gen cards. I would not expect more unless we get newer info. As always, I am hoping I am wrong and you are right, that the 4870 will be mindblowing, but go back to my first sentence and you will know why it is good to be cautious.
Perkam
perkam that's highly unlikely and you know that. The worst scenario we've seen for the 4870 is gddr5 memory with 480 shaders and more TMUs. I'd be willing to bet the 480 shaders is correct because that would support the rumors about the 4470 and 4670 with 120 and 240 shaders (meaning clusters of 120). TMU number is beyond me though if we're going to clusters of 120, but it's definitely going to be higher, and there's a lot of talk of 3:1 Alu:TMU with 96:32.
So worst come worst, we'll see at least a 50% gain due to the 50% more shaders and higher clocks, meaning at least a good 10% over the 9800gtx. And that's not accounting for how performance will be without the huge TMU bottleneck and plenty of shaders for AA
I definitely can't wait to see tangible results instead of all this speculation.
EVERYBODY and their mothers already know that the 8800 GT is a better card then the 3870!!! The 3870 isnt meant to be better, it always was slower then the 8800GT.
And where exactly is your past experience coming from?
ATI 8500 was absolutely equal to, if not faster then the Geforce 3.
ATI 9700 / 9800 owned the entire Geforce 5 range, the geforce 5 being the WORST GPU the industry has ever seen.
X800 / X850 were on par with the 6800's, but they lacked DX9.0c
X1900 - X1950 were again faster then everything from the Geforce 7 range, particularly in DX9.0 intensive games.
ATI were screwed over when they were aquired by AMD, and the only single launch that was messed up and delayed since the dawn of the first Radeon chip was the 2900, however both the 2900 and 3800 are at least capable of keeping up with Nvidias current cards, and the 3800's scale far better when put into crossfire then Nvidia manages with SLI.
ATI, in their entire history of graphics cards, have IMO messed up one single product launch, and that was the 2900. And it wasnt their fault, it was the aquisition by AMD that somehow slowed them down.
Were you even around when the Geforce 5 was out, or aware of how :banana::banana::banana::banana: it was? ATI have never created a graphics card yet that is anywhere near as bad as the geforce 5 was.
The 4800's should be a very strong comeback for ATI, and hopefully they wont mess up any launches again. It is a much bigger improvement over the 3800 then you think, you just seem to be grossly underestimating what the card will be capable of.
Another Nvidiot right here that has nothing constructive to add.
The 9800 is a far more powerful GPU compared to the 3870, comparing those two together is something an Nvidiot would do (omg 9800 > 3870, Nvidia Ruuuuuuullleeeeessss!!!)
If you compare two 3870's to two 9600 GT's in Crossfire / SLI, you will see that the 3870's gain a lot more performance then their Nvidia equivalents.
And yes, I said SCALE not Outperform.
If you put two 3870's in crossfire, you get a much higher %age boost then you would from putting two Nvidia cards in SLI.
Oh yea, I forget about GPU's from companies from anyone other then ATI / Nvidia. I dont really care about them :p
ATI X1800, hmmmm, I dont think it was as late as the 2900 was, but it was quickly replaced by the superior X1900 anyway.
And Nvidia are actually currently late with the 'real' Geforce 9. The current 9800 range is just a cover up for the fact that they couldnt get the actual DX10.1 chip out on time, but they coverd it up o so well by renaming the 8800's :p
still the x1800 was a decent card at launch, it pushed nv to the "HAHA YOU'LL NEVER GET US 7800GTX512" not available in any store :p:
ati had the upper hand from 8500-x1900 (nv GF3-GF7 series), AMD put a stop to this, but things are looking good that they come back on top (they got back in the game with the 3870s)
to point 1: we talk about dedicated graphics cards, and this title belongs to the rage fury maxx :D, but in this times we had no shaders :shrug:
hd4870 looks promising, it may not beat the gt-200, but this time ati is earlier to the market and i believe that the gap between amd and nv is going to get smaller.
Yes, I know the 8800GT is a better card as well, and I already stated the 4870 will a very good attempt to catch up, so I dont get how I was wrong there. Well, if you dont like me calling it past experience you can call it something else. But whatever it is, it gives you the intuition of knowing that if a 320-shader, 256-bit, 16/16 TMU/ROP card (HD 3850/3870) only manages to get twice the 3d06 points compared to a 120-shader, 128-bit, 8/4 TMU/ROP card (HD 3650) on the same architecture, a 160 shader increase with no change to the 256-bit bus and no change to ROPs will not have a performance effect greater than 30% at best (my 12.5k estimate was a 25% increase over the current 3870s stock 10k score).
That was my logic. It is embedded in realism, not wishful thinking.
Perkam
We still dont know if the shaders have their own clock domain. If that's true your calculations are wrong. Also, 3870 is bottlenecked by its 16TMUs, you don't know how 32TMUs can change things. My bet is 14k minimum. And a 3870 scoring 10k? With what CPU?
The X1800 were good cards (I had 2 of them) and i'm not gonna argue that but the launch was what went wrong. ATi did get some stick from X1800 purchasers for the release of the X1900's but that soon waned and was replaced by cheers from the majority.
GF3 was actually faster than the 8500 in OGL & DX7 games but not synthetics but that's not the point. Infact it was almost a whitewash for the GF3 until ATi fixed their drivers. It had better features but lacked the proper software to unleash it.
Early Crossfire was a bit haphazard aswell but if you look at where they are now, they have once again got a advantage in API (questionable benefit tyet), process (55nm) & have almost made CF an excellent product (profiling still needs to be opened up more) but they still make the odd little mixup along the way (not enough texturing power in R600/RV6x0).
HD 4xxx is again looking at keeping what works intact (with a little improvement) but fixing what they went wrong with before. While it's not looking like enough to retake the performance crown it might enable them to increase the prices enough to actually make a profit, which lets face it, they could really use.
I still see the Intel 740 as a dire product, after all it is the grandfather of the GMA series...
You dont get it :)
If you get 10k stock currently with the 3870, I'm predicting you'll get 12.5 - 13k stock with a 4870.
If you get 11k stock currently with a 3870, I'm predicting you'll get 12.75 - 13.25k stock with a 4870.
If you get 12k stock currently with a 3870, I'm predicting you'll get 13 - 13.5k stock with a 4870.
As I said, 25-30% is realistic. Any more than that is wishful thinking due to the reasons provided in my last post here:
And you can't discount the reasoning there. If the 3650 was 256-bit the score would be around 6.5k, which would make the 3870 score with almost 300% the shaders (and double the TMUs AND triple the ROPS !!) of only 10k completely incompetent, so you can see only 50% more shaders will not have a huge effect. It's called diminishing returns, look it up.Quote:
Originally Posted by Perkam
I have my reasoning for my estimated score, I would like to see yours.
Perkam
Imo your reasoning is flawed here. There is no way to predict the effect of a faster bus on the 3670. There is no way you can know whether its lack of TMU/ROPs is going to allow it to have anything more than minimal gains from a faster bus. Additionally what the 4870 should bring is improved real world gaming performance as it seems to address what appears to be the biggest bottleneck in the R600 architecture for many games