No Fermi benchmarks / Price & TDP revealed

[tviceman]

And now added to the latest string of mystery updates from the same guy at VR zone who seems to have all the info, the TDP has been lowered from 275w to 250w.

http://forums.vr-zone.com/7792771-post44.html

 

[BFG10K]

Whatever the TDP is, I hope those grills on the side don't allow air to escape into the case.

 

[Keysplayr]

Well, we've survived that travesty before. No doubt we can survive if again if it is the case (no pun intended).

 

[dguy6789]

I scanned that thread and, in my opinion, it wasn't nearly as negative or slanted against AMD as threads have been about nvidia lately. Not even close.

We have Nvidia's next gen architecture launch right around the corner. A new video card generation launch is probably the most exciting and busy time for people on this forum. A little hysteria should be expected me thinks. It doesn't matter what Nvidia produces, some people won't be happy. The same goes for AMD. It can't be helped.

 

[Lonyo]

We have Nvidia's next gen architecture launch right around the corner. A new video card generation launch is probably the most exciting and busy time for people on this forum. A little hysteria should be expected me thinks. It doesn't matter what Nvidia produces, some people won't be happy. The same goes for AMD. It can't be helped.

Plus at that time prices were just staying the same, and there was competition further down the ranks.
This time the market is stagnant pretty much from top to bottom, except where prices have gone up. It's a much worse situation for the consumer.

 

[n7]

They better be a hell of a lot faster than i think they will...or we're gonna have prices going up on HD 5800s. :\

This was precisely my fear, which is why i ordered my HD 5870 last week.

 

[dguy6789]

The video forum these days is like a mini P&N section. We have party loyalty, crazy arguments over virtually anything, and crazy journalists.

Hey look it's Charlie!

http://www.youtube.com/watch?v=H-D_S7WOnjg

:sneaky:

 

[jaggerwild]

At least not on the 26th.
http://www.nordichardware.com/en/co...vidia-moving-media-embargo-to-march-29th.html


Prices:

GeForce GTX 480 : 512 SP, 384-bit, 295W TDP, US$499

GeForce GTX 470 : 448 SP, 320-bit, 225W TDP, US$349

http://vr-zone.com/articles/nvidia-geforce-gtx-480-final-specs--pricing-revealed/8635.html

In the link that photo you can clearly see the pallet of boxes waiting to ship but the Nvidia card box isn't the same box, Fail!

 

[MrK6]

At least not on the 26th.
http://www.nordichardware.com/en/co...vidia-moving-media-embargo-to-march-29th.html




Prices:

GeForce GTX 480 : 512 SP, 384-bit, 295W TDP, US$499

GeForce GTX 470 : 448 SP, 320-bit, 225W TDP, US$349

http://vr-zone.com/articles/nvidia-geforce-gtx-480-final-specs--pricing-revealed/8635.html

Considering I just paid $510 for 5850 Crossfire, which is more or less the same TDP as a GTX480 and completely destroys it, wow, good luck with that one NVIDIA.

 

[McWatt]

http://vr-zone.com/articles/nvidia-geforce-gtx-480-final-specs--pricing-revealed/8635.html

GeForce GTX 480 : 480 SP, 700/1401/1848MHz core/shader/mem, 384-bit, 1536MB, 250W TDP, US$499

If this is true, and corroborating stories are showing up on places like fudzilla now, then Charlie has been caught pretty badly this time. Less than a week ago he was claiming 275 W, and he's insisted that the core clock would be 600-625 MHz, off of a 750 MHz target.

It looks like they missed their target by 6.7%, not the 20% Charlie expected. His reporting of how much clocks had slipped was 3x the actual value. In the future, perhaps his articles should be taken with a factor of 3 grain of salt. While his delay reports seem to have been vague but generally true, his spec guesses seem worse than just rolling dice or picking random numbers.

Now a question: do nVidia cards support audio over HDMI? If not, clocks and TDP won't matter to me.

 

[scooterlibby]

http://vr-zone.com/articles/nvidia-geforce-gtx-480-final-specs--pricing-revealed/8635.html

GeForce GTX 480 : 480 SP, 700/1401/1848MHz core/shader/mem, 384-bit, 1536MB, 250W TDP, US$499

If this is true, and corroborating stories are showing up on places like fudzilla now, then Charlie has been caught pretty badly this time. Less than a week ago he was claiming 275 W, and he's insisted that the core clock would be 600-625 MHz, off of a 750 MHz target.

It looks like they missed their target by 6.7%, not the 20% Charlie expected. His reporting of how much clocks had slipped was 3x the actual value. In the future, perhaps his articles should be taken with a factor of 3 grain of salt. While his delay reports seem to have been vague but generally true, his spec guesses seem worse than just rolling dice or picking random numbers.

Now a question: do nVidia cards support audio over HDMI? If not, clocks and TDP won't matter to me.

I seem to remember that target clock was never claimed by Nvidia, rather it was suggested by AMD as a PR highball move - can someone else back me up or refute this?

 

[Daedalus685]

http://vr-zone.com/articles/nvidia-geforce-gtx-480-final-specs--pricing-revealed/8635.html

GeForce GTX 480 : 480 SP, 700/1401/1848MHz core/shader/mem, 384-bit, 1536MB, 250W TDP, US$499

If this is true, and corroborating stories are showing up on places like fudzilla now, then Charlie has been caught pretty badly this time. Less than a week ago he was claiming 275 W, and he's insisted that the core clock would be 600-625 MHz, off of a 750 MHz target.

It looks like they missed their target by 6.7%, not the 20% Charlie expected. His reporting of how much clocks had slipped was 3x the actual value. In the future, perhaps his articles should be taken with a factor of 3 grain of salt. While his delay reports seem to have been vague but generally true, his spec guesses seem worse than just rolling dice or picking random numbers.

Now a question: do nVidia cards support audio over HDMI? If not, clocks and TDP won't matter to me.

Charlie was claiming a miss for the 512 core gtx480... Does it matter who was wrong if no one was right? Hell, if they released it with even less cores and surpassed their target does that make Charlie even more retarded or does it just become absolutely silly?

 

[Stoneburner]

so the gf100 will be 512sp?

 

[Daedalus685]

so the gf100 will be 512sp?

The GTX480 is rumoured to be only 480, which seems very likely... Though many in the media are still of the impression that a 512 core card will follow at some point.

The gf100 chip itself is 512 cores though, yes.

 

[crisium]

GTX 490 with 512 core? Might make sense if the rumours of "Fermi too big for dual-chip card" are true.

 

[McWatt]

Charlie was claiming a miss for the 512 core gtx480... Does it matter who was wrong if no one was right? Hell, if they released it with even less cores and surpassed their target does that make Charlie even more retarded or does it just become absolutely silly?

I think it just further demonstrates that Charlie's information isn't very solid. I'm not on Team Red (though I've had an ATI for the last several generations), Team Green (though I think nVidia is fine, too), Team Edward or Team Jacob. I don't have an agenda and I can't even understand the people who politicize graphics card companies. I just want to know how much I can trust the numbers produced by the early rumor sites like semiaccurate, fudzilla, etc. The answer seems to be "not much."

 

[tviceman]

Now a question: do nVidia cards support audio over HDMI? If not, clocks and TDP won't matter to me.

My gtx260 has spdif out and a dvi output, so add in a $10 hdmi-dvi adapter and a gtx260 and "yes."

At the bare minimum, I don't see fermi being any different. Then again, they may have had to cut audio out of the high end for size reasons - I don't know.

 

[RussianSensation]

Chances are he received vague information on the specs. Technically GTX 470's clocks fall between 600-625, which is way under 725 of the highest model with 512 SPs. But it was probably said "The cards are clocking at 600+, not 750". Then, he spun that to mean the highest model (which is incorrect).

The best part here is that GTX 470's specs aren't cut down that much compared to the 480 SP model. I bet factory overclocked versions will be able to match stock 480's performance and still save one $100 over GTX 480

I am still struggling to understand how with nearly double the SPs over GTX 285, the GTX 470 is only 10% faster than 5850. I think with driver improvements, it will match 5870 (hopefully).

 

[Stoneburner]

I think it just further demonstrates that Charlie's information isn't very solid. I'm not on Team Red (though I've had an ATI for the last several generations), Team Green (though I think nVidia is fine, too), Team Edward or Team Jacob. I don't have an agenda and I can't even understand the people who politicize graphics card companies. I just want to know how much I can trust the numbers produced by the early rumor sites like semiaccurate, fudzilla, etc. The answer seems to be "not much."

This presupposes that the facts as they currently exist were available back then. My impression is that Nvidia didn't finalize specs until a short time ago and that the sites were getting information regarding the product that was current at the time it was published.

 

[scooterlibby]

GTX 490 with 512 core? Might make sense if the rumours of "Fermi too big for dual-chip card" are true.

I was completely confused by the Vr-Zone quote : "The intended GF100 has 512 SP clocked at 725/1450/1050MHz with 295W TDP. It should still be released in the future but just not now. "

I guess maybe a 490 may be what they are implying.

 

[Udgnim]

pretty sure Charlie said that any Fermi chips that had 512 cores were going to be reserved for Nvidia's Tesla line

 

[tviceman]

I am still struggling to understand how with nearly double the SPs over GTX 285, the GTX 470 is only 10% faster than 5850. I think with driver improvements, it will match 5870 (hopefully).

Perhaps driver improvements between leaked benchmarks and official reviews will reflect that. Either way, while benchmarks of dx9/dx10 games are good and needed for comparision purposes I am most interested in seeing how performance is in DX11 games - aka most future releases and fermi's supposed bread & butter.

 

[thilanliyan]

This TDP obsession is really retarded.
My GTX285 OC has a TDDP of ~250 Watt.
If Fermi has a TDP of 295 Watt...but ~2x the perfomance...why should I care?

Unless I am from AMD's camp, worried about performance so I start focusing on abitary sticks...

Some people DO care and rightly so. I want to keep tdp down not because of my PSU (HX620w) but because of the cooling. I like mid-high end components but also like things quiet so I use water cooling. However, like air cooling, there's only so much heat a WC system can dissipate and if Fermi by itself is putting out close to 300w, that would definitely not bode well for my system (2x120mm + 1x120mm radiators with low speed fans). All things being close to equal, I'd take the card with lower tdp even if I get slightly lower performance.

 

[SlowSpyder]

http://vr-zone.com/articles/nvidia-geforce-gtx-480-final-specs--pricing-revealed/8635.html

GeForce GTX 480 : 480 SP, 700/1401/1848MHz core/shader/mem, 384-bit, 1536MB, 250W TDP, US$499

If this is true, and corroborating stories are showing up on places like fudzilla now, then Charlie has been caught pretty badly this time. Less than a week ago he was claiming 275 W, and he's insisted that the core clock would be 600-625 MHz, off of a 750 MHz target.

It looks like they missed their target by 6.7%, not the 20% Charlie expected. His reporting of how much clocks had slipped was 3x the actual value. In the future, perhaps his articles should be taken with a factor of 3 grain of salt. While his delay reports seem to have been vague but generally true, his spec guesses seem worse than just rolling dice or picking random numbers.

Now a question: do nVidia cards support audio over HDMI? If not, clocks and TDP won't matter to me.

I don't think Charlie was as wrong as you're making it sound. I think in a number of stories (not just Charlie, but other sites as well) he mentioned that not even Nvidia had the final specs nailed down, some sites had stories as recent of a few days ago saying the final specs may not have yet been set in stone. So, with a 512 SP part Charlie was probably correct in what he posted, but by Nvidia moving to a 480 SP part with a 700MHz clock they ended up at the 250 watt TDP part. I'm guessing that the 512 SP part may have been closer to the TDP/clocks he reported.

It seems to me that Charlie had many pieces of the puzzle, but not everything obviously. That's why it's good to take rumor site stories with a grain of salt, a lot of what they post is based on fact, but obviously things change and what they post may only be a portion of the big picture. I think that's exactly what happened here... the information was most likely very much correct, but there was more information to be had to get a truely accurate idea of what what going on.

 

[Soleron]

They can give the GTX 480 any clockspeed they like. It won't change the fact that none exist. If you can't produce any at 1200MHz and you can't produce any at 1400MHz you might as well spec a 1400MHz part and get the 3 out of 10000 that do it to reviewers.

The thing that Charlie was sure of was that no part will appear in volume above about 1200MHz. Beyond that, he said the samples that he received numbers from were 1250MHz clocked. As we saw, final specs were not known when he wrote that article, he was reporting the specific sample of the 470 and 480 he knew about.