Early 470 & 480 benches

[cliffmidnite]

well, all questions should be answered within 10 days when the cards are unveiled at PAX. I just read today that the official retail release date is pushed back till April 6th. I cant wait to see some real unbiased gaming benchmarks and what the retail price is going to be. I don't see anyway Nvidia is going to win on Performance per dollar.

[bobzilla2009]

I don't see where they got the hd5870 results for heaven from (in the OP), since mine did better than that at stock! I get around 35fps at those settings with my OC too. So my OC'd card beats the gtx470 (if it is that card) by ~20% on it's home field benchmark (heavy tessellation). Win!

[johnnyfiive]

Heh, I'm becoming inpatient and will soon have plenty of money to spend. I could get a 58xx right now and be a happy johnnyfiive, or wait for fermi and be an inpatient, probably disappointed johnnyfiive when prices are revealed and then be screwed because everyone waiting bought up all the 58xx cards I could have gotten.

[bobzilla2009]

Originally Posted by johnnyfiive

Heh, I'm becoming inpatient and will soon have plenty of money to spend. I could get a 58xx right now and be a happy johnnyfiive, or wait for fermi and be an inpatient, probably disappointed johnnyfiive when prices are revealed and then be screwed because everyone waiting bought up all the 58xx cards I could have gotten.

hd58xx prices will probably drop anyway when the fermi cards come out. So you can always wait, it's not like it's 6 months away until fermi (like when i bought my card).

[3volvedcombat]

Originally Posted by Steevo

http://i46.tinypic.com/2a8ncsz.jpg

Nvidia offering.......



Mine.

http://img.techpowerup.org/100314/heaven.jpg

http://www.techpowerup.com/gpuz/hhxxw/

Thats a GTX 470 because i can see 1200mb of memory which the 470 has, also you have your Phenom and rig tweaked up to 3.8Ghz. With 1337 mem sets.

Its barely a stock clocked i7 920 at 3.3Ghz they tested with, and the Video card benching scores they showed were on beta bios'ed cards and beta designs that arnt even overclocked. They Got that score purely stock clocked on the processor(Intel Speed Step) and everything. So unless your dreaming your 3 fps from getting BUT raped by a 470 and a stock clocked i7 lmao....

[bobzilla2009]

Originally Posted by 3volvedcombat

Thats a GTX 470 because i can see 1200mb of memory which the 470 has, also you have your Phenom and rig tweaked up to 3.8Ghz. With 1337 mem sets.

Its barely a stock clocked i7 920 at 3.3Ghz they tested with, and the Video card benching scores they showed were on beta bios'ed cards and beta designs that arnt even overclocked. They Got that score purely stock clocked on the processor(Intel Speed Step) and everything. So unless your dreaming your 3 fps from getting BUT raped by a 470 and a stock clocked i7 lmao....

heaven is nvidias best benchmark on fermi though. It's the only one where they can decisively win over the hd5 series due to the huge amount of tessellation. The fact that the difference between gtx470/80 isn't a lot and the gtx470 still can't beat the average frames of a hd5870 (even when overclocked) on nvidias very very best benchmark spells doom. Do you think a gtx470 will keep up with a hd5870 in a real game if this is the case?

Also, just to point out, heaven is not remotely cpu limited, the processor won't make any difference. it's purely cards. Anyway, i'm sure no hd5870 owner is going to lose sleep if the gtx480 wins by 5%-10% if they've owned the card for nearly 6 months (i'm certainly not going to). Btw that result for heaven for the hd5870 seems a little low, i got 35.7 fps at the same settings (1920x1200, 4xAA and approximatley the same OC [a little lower in my case, 'only' 1040MHz) :S

[dr emulator (madmax)]

well i wonder how many old classic games will run on this new tech, without the need for big patches, only time will tell

[Blacklash]

Heaven on my system does 35.4 FPS @ 1920x1200 stock default filtering. Overclocking my 5870 raises it to 41FPS. Of course that's DX11. DX10 it will do 66FPS @ 1920x1200 default filtering.

BTW if you want a 10.1 bench from the same guys try their tropics bench below.

http://unigine.com/download/

[Steevo]

Originally Posted by 3volvedcombat

Thats a GTX 470 because i can see 1200mb of memory which the 470 has, also you have your Phenom and rig tweaked up to 3.8Ghz. With 1337 mem sets.

Its barely a stock clocked i7 920 at 3.3Ghz they tested with, and the Video card benching scores they showed were on beta bios'ed cards and beta designs that arnt even overclocked. They Got that score purely stock clocked on the processor(Intel Speed Step) and everything. So unless your dreaming your 3 fps from getting BUT raped by a 470 and a stock clocked i7 lmao....

The clock effiency makes at least the difference between my Phenom and a i7. Plus as was pointed out the CPU makes jack all for difference in what I have tried in Heaven benchmark, the whole idea of DX11 is to remove the CPU from the equation as much as possible.


Plus a 470 is officially 1280Mb, not 1200. So it is a FAIL, beta could also mean they had the clocks ramped up and were phase cooling it, there are no temps shown, no clocks shown, nothing. So in all fairness, I shouldn't show clocks, temps or otherwise. Just pure card on card hot sexy benching action. And my measly 1Gb 5870 HIS overclocks 24% on air out of the box, for all of $379, is available, and came with a free game, and has a much lower TDP and probably temps, and quieter.


Tell me your sorry story when you actually have a card in hand.

[bobzilla2009]

Originally Posted by Steevo

The clock effiency makes at least the difference between my Phenom and a i7. Plus as was pointed out the CPU makes jack all for difference in what I have tried in Heaven benchmark, the whole idea of DX11 is to remove the CPU from the equation as much as possible.


Plus a 470 is officially 1280Mb, not 1200. So it is a FAIL, beta could also mean they had the clocks ramped up and were phase cooling it, there are no temps shown, no clocks shown, nothing. So in all fairness, I shouldn't show clocks, temps or otherwise. Just pure card on card hot sexy benching action. And my measly 1Gb 5870 HIS overclocks 24% on air out of the box, for all of $379, is available, and came with a free game, and has a much lower TDP and probably temps, and quieter.


Tell me your sorry story when you actually have a card in hand.

tbh it does look like there is an 8 in the memory shown on the heaven pic, but it's blocked by a stupid symbol, as it the card number too (although that does look like a 7). But it doesn't matter anyway, fermi is going to be a nice, average but overpriced card. Nvidia fans have waited for ages listening to how the card was going to pwn the hd5870 harder than a hard thing on a hard stick, just to discover it's actually pretty much on par. Can you imagine how that must feel, waiting all that time seeing all those happy hd5870 owners run crysis on very high with lots of lovely AA? waiting 6-7 months for a card that's only as powerful as a card that was available 6 months ago! So i suppose the nvidia fanboys deserve a break, at least for a little while. let them man up to take it on the chin from their beloved card bearing diety, it's going to hurt.

P.S. I'm just a little fed up of the whole fanboy mentality atm lol. Just had a 'debate' about fermi (I'm annoyed at the fanboys bigging up a card that is basically the same as the ATi card from yesteryear) until the fanboy broke, realizing impending heat/power draw/performance-price related defeat and started spouting 'But physX is awesome! CUDA pwns all!' I couldn't be bothered, the barrel is scraped when those arguments appear. Proprietary tech does not make a card, especially when the guys that are shut out are the ones that are rapidly gaining market share in the gaming market:

The most owned DX10 cards are the hd48xx series of the last generation by a long way
The biggest change in % ownership in the last 5 months have been the hd58xx and hd57xx series cards.
Nearly all of the 65% nvidia market share is made up of the old 8 series and older cards (or their 9/2 series derivatives).

If i were a game developer and looked at these stats, i would say ATi is the way to go as most gamers with the best kit have ATi cards. Thats what developers have done. The amount of twimtbp games has nosedived heavily since last september.

EDIT: made my point clearer, i forgot to mention that i only included the last gen for the hd4 stat, of which it wins by a long long way. If you also count the 8800 and the 9800 as different cards (which is wrong, although nvidia got sales by saying just that), the point also holds true.

[HalfAHertz]

That's bull, the most owned Dx10 card is the 8800 and its derivatives(how many renames has it been so far 3 or 4?) :P

[KainXS]

I agree its bull the most owned DX10 card series is without a doubt the 8800 series, the 8800GT notably.

the most owned gpu's are the G92's.

[bobzilla2009]

including derivatives i have no doubt it's the 8800 , what i should of put was 'and not including cards that are just copies' ^^ i've edited the post to make my points a bit clearer

[alexsubri]

I just can't wait to see how it's all going to unfold. nVidia has a lot of answering to do and everybody is waiting to see if the new Fermi card's can deliver. I presonally can't wait to see REAL reviews and benchmarks on the card. So as of now, I'm still meh with the whole Fermi ordeal.

[nt300]

Kiss Fermi's lifespan bye bye
Nvidia in big trouble, this card runs at 98C in less than a minute in Future Mark benchmark. The new Fermi's look like a very short life span. I would stay away from these cards. The Article has confirmed sources

How te Hell are you going to keep your game PC cool with a oven hot GTX 480 & 470

Nvidia bends the definition of honesty in GTX480 benches!
http://www.semiaccurate.com/2010/03/...tx480-benches/

Nvidia bends the definition of honesty in GTX480 benches
Same old same old, but this time much hotter!
by Charlie Demerjian
March 23, 2010

NVIDIA HAS SOME interesting numbers in its GTX480 presentations, but just like the Heaven benchmark numbers, they don't seem to reflect reality. This latest 'accident' in an official presentation centers around Dirt2.

When SemiAccurate put up the first set of GTX480 benchmark numbers a month or so ago, people laughed at our claims of a small margin over a 5870 while losing badly to a 5970. The fanbois were almost apoplectic over our claims of 70C idle temps and far hotter in gaming. People with fragile egos, especially those tacked to dreams of enthusiast hardware just hate having their bubbles burst.

Then came the Nvidia benchmark snippets. First was Heaven, or the parts of it, strangely only the parts where Nvidia did well. From there, bits and pieces trickled out, all cherry picked. One that was featured in official Nvidia presentations was Dirt2, the 'ATI centric' benchmark. For some odd reason, Nvidia's GTX480 won handily, it won by so much that something seemed very odd.

As it turns out, the numbers were indeed too good to be true. If you run the Dirt2 demo, the GTX470 and GTX480 drop back to DX9 mode, as you can see it in the picture above. Since they are doing far less work, frame rates go up. Running the DX11 code path drop frames by around 25 to 40 percent for the same work. In the game itself, DX11 works just fine on the GTX480, so it is likely that the demo lacked the correct profile for the then unreleased and 6 months out GTX4x0s.

For some odd reason, that point wasn't mentioned in the Nvidia slides SemiAccurate saw. Sources deep inside Santa Clara have told SemiAccurate that this wasn't due to the TWIMTBP budget cuts, it is probably just the old 'Nvidia honesty' coming forward once again. For some reason, the real numbers that compare DX11 to DX11 versions didn't make their press presentations even though the game had been out for months by then. Funny that.

For those unwilling to take such things as hard facts into account in order to protect their egos, we submit the above screenshot of the Dirt2 demo, running on a GTX480. The card is running at the official 1848 MHz (3696MHz effective) memory clock and the shaders are at the stock 1401MHz 'hot clock'.

Please note, in this case 'hot' is more literal than figurative, the GPU here is running at 87C, far hotter than any card that expects to have a realistic life span should be at. Unconfirmed reports from China have the card hitting 98C on furmark. Don't expect these 'puppies' to have a long life, even in dog years.

In the end, the card is too hot, too slow, and unmanufacturable. We told you so. Pop goes the ego.S|A

[sneekypeet]

furmark, not futuremark.

And who tests a card on a demo?

[erocker]

Yep, happens to HD 4870's too. Furmark is a lousy program. Also, I found a thread this topic fits in, as Fermi doesn't need any more threads. 2 more days.

[RejZoR]

No way. A whooping 98°C ? Are you guys kidding or what? That's a perfectly normal thermal level for almost any NVIDIA GPU. In fact most of them run perfectly fine up to around 115-118°C. Over that it starts to throttle and the fan goes to 100%. And we all know hardly anything tortures GPU's as much as FurMark. In fact, the only thing that comes half that close is Crysis on Very High with FSAA and AF. This was the only game where my fan started spinning faster than usual, but not nearly as fast as with FurMark. So in normal games, it wouldn't even reach those 98°C.

[EastCoasthandle]

Originally Posted by RejZoR

No way. A whooping 98°C ? Are you guys kidding or what? That's a perfectly normal thermal level for almost any NVIDIA GPU. In fact most of them run perfectly fine up to around 115-118°C. Over that it starts to throttle and the fan goes to 100%. And we all know hardly anything tortures GPU's as much as FurMark. In fact, the only thing that comes half that close is Crysis on Very High with FSAA and AF. This was the only game where my fan started spinning faster than usual, but not nearly as fast as with FurMark. So in normal games, it wouldn't even reach those 98°C.

I think you've forgotten how radiant heat from 98°C would effect other thermal sensitive components near the video card. The gist of it is that you would need an above average ventilated PC Case. Something everyone doesn't have or need to buy.

[Binge]

that may have been a good temp for GPUs on a larger process but they're in a serious danger zone for the component itself. If you want to get even more to the topic, the solder they use to attach the IHS to the GPU core melts at around 95C. Imagine if it just lets go of the IHS one day?

[RejZoR]

Originally Posted by EastCoasthandle

I think you've forgotten how radiant heat from 98°C would effect other thermal sensitive components near the video card. The gist of it is that you would need an above average ventilated PC Case. Something everyone doesn't have or need to buy.

I'm using a uATX factor case and inside, there is a Core i7 920 and HD4870 sweating with nearly no air circulation because i hate jet engines. I've even replaced a fan inside PSU to make it fully silent. And that's the one pumping hot CPU air from the Core i7.
HD4870 is running on lowest possible RPM up till 75°C where it slowly goes to 100% (at 100°C).
And it feels perfectly happy with that and hardly ever speeds up much so it's almost silent even in silent moments inside games. So, if you take this into account, pack a graphic card in a usually well ventilated and reasonably big case, there is no problem. And considering NVDIA is favoring "the duster" coolers where all the heat is dumped outside (where my HD4870 is dumping half the heat inside the case and other half outside), there shouldn't be much problem really.

Besides, looking at this from a life expectancy view is really pointless. Especially with high end cards where they are usually swapped soon, far sooner before they'll ever crap out.
If you have cash for such card, it's a very big chance you'll soon swap it for another very high end model. Where lower end cards serve for longer periods because they are generally crappy to begin with and are used by non demanding users anyway. Pretty much all have 2 year warranty as well, so in 2 year time frame, you'll change it either way or get a new one with warranty. So all is well if you ask me.

[KainXS]

98C on load and 70C idle, hmmmmm, with a 1.8A fan, . . . . . . . . . well . . . . . . . . if that is true then fermi is fail but since its charlie, ima wait for a real benchmark instead of believing "him".

98C after 1 minute of furmark is not good, I had a old GTX with a bad fan that hit 105(it was more than likely much higher), then the sensors on it failed, and read 255 no matter what, like a day later the card itself stopped working.

[Lionheart]

I thought semi accurate was a dud site.

[GC_PaNzerFIN]

http://www.news4it.it/gpu/175-nvidia-gtx480-3-way-sli

[erixx]

Hi ALL!!!!
I just registered after reading for some years (i know: bad, bad, wrong! hehe)

Just wanted to state my surprise reading those 80-98 temps!!!

I have a pair of SLI'ed XFX 9600 XXX (factory overclocked) in a well vented case (LianYi PC-70 full tower) and my temps are 40 for idle and 60 for stress.

I don't like high temps at all.... (and I am seriously considering jumping onto a 5850 soon for a better energy profile and better performance)

Well, in short, 98 degrees is crazy!