GF100 (Fermi) previews and discussion

Page 26 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Toggle sidebar Toggle sidebar
Status
Not open for further replies.
Thing is, all A2s arent slower than A3s, its just most are heheh.
If it cuts the mustard, I say go for it, they did, sorta like AMD triples, maybe theyll go quad, and a sales a sale, and if numbers are as low as feared, again, why not?
 

It seems to be working for ATi with the broken 5850/5870 chips being flogged off as the 5830, heh heh.
 
I guess people at Quality Control finally reached an agreement with the people at marketing, have a defective chip lower the price, disable parts to be more uniform and call it a new chip!
 

It also makes it easier to extrapolate the issues that Nvidia were going to have if over the course of a few months you were to see several skips filling up with defective GPU's that have an ATi logo on them, I saw an article a while back that suggested these chips have been getting slung out since August which makes me wonder if that's what Charlie has been basing his info on.
 
Remember, that both will have the via issue, it's just how you anticipate it.

40% yield with 50% bad is still better than 10-20% yields without the 'fall back' designs.

It wasn't just charlie and Anand mentioning this, this was brought up in December when yields were 'really low' and still they got product out, but likely a ton of dies that were just unusable. Even at a good (by old standards) 70% yield, if you're selling 3 milion+ GPUs that means 1.2 million+ more are sitting in a barrel somewhere. :??:

Seeing both A2 and A3 in what amounts to a hnadful of select Cebit boards doesn't say much of anything, just that A2 was usuable, and so is A3, in what quantities likely no one other than TSMC and nV know for now.
 
GTX 480 to have a 300W TDP

http://www.fudzilla.com/content/view/17921/1/

brrrr, if thats true then its even worse than expected. Smell of baked plastic anyone :)
 


Hey man, Charlie has had an average 6 month track record based on extremely predictable events, he deserves more credit despite his years and years of bad information before that! The Fermi isn't hot, broken, and unfixable? This must be FUD, I suggest we all wait for Charlie's inside info before jumping to any conclusions.
















/endsarcasm
 


Well, technically they could have invented some amazing cooling device for the card. The card could still be hot as hell 😉. I myself am really excite to see what they cool a 300W single GPU with.
 
http://translate.google.com/translate?js=y&prev=_t&hl=de&ie=UTF-8&layout=1&eotf=1&u=http%3A%2F%2Fwww.heise.de%2Fnewsticker%2Fmeldung%2FNvidias-Fermi-Leistung-der-GeForce-GTX-470-enthuellt-946411.html&sl=de&tl=en


Vague and poorly translated performance values.

5850<470<5870 almost across the board. The 470 wins in unigine with low levels of AA, but can't keep up with 8x. This, if true, seems to confirm the idea that the fermi is only better at tessellation when it doesn't need to cores for other computation.

We will see.
 
^^ From what I read, the 470 was about equal with the 5870, even at the 8x AA level. The 5870 is just a smidge ahead based on those numbers. [+157 points in 3dmark Performance mode, +4 FPS in Unigine at 8x AA]. And given the 480 has an extra bunch of shaders, the 480 *should* be significantly faster then the 5870.

Encouraging first numbers if true.
 
Even if those numbers were true they offer absolutely no indication of how they will perform in games or how much they will cost. We want independant benchies dammit. Of course this doesn't affect me but it would shut people up one way or the other at least.
 


Well, that is one way to spin it I suppose. But given the praise for fermi to save the world I must say if this all turns out to be true I will be disappointed in the dx11 results. I agree that 4fps is almost nothing... but we are talking about 23fps for the fastest card there, and the difference between the 5850 and 5870. (Fermi is actually 3 fps slower in that test, not 4)

3 or 4 FPS works out to between 10-20% in a case like that.. Other benches will very wildly, but that is not the double unigine values we have heard before, unless the 480 is more than twice as fast as the 470..

I'd have to assume we are going to see something akin to the 4870 1gb vs the 260 core 196 and on the highest end it might be equivalent to the gtx280. Though right now I'm sceptical the 480 will be as much faster compared to a 5870 as a 280 was to the 4870. Hopefully I am wrong.

Right now we are looking at what, 15% more shaders for the 480? Given linear scaling it would only tie the 5870 in the AA heavy unigine. Oh well, random translations have failed us before.. I still have my fingers crossed that I will be pleasantly surprised, though I'd like to get some new games to use even my old tech fully...
 
Not looking that geat if this is true
http://www.heise.de/newsticker/meldung/Nvidias-Fermi-Leistung-der-GeForce-GTX-470-enthuellt-946411.html
3DMark Vantage, X-Mode

RadeonHD5850 6430
GeForceGTX470 7511
RadeonHD5870 8730

3DMark Vantage, Performance Mode

RadeonHD5850 14300
GeForceGTX470 17156
RadeonHD5870 17303

Unigine tesselation, 4x AA

RadeonHD5850 22 fps
GeForceGTX470 29 fps
RadeonHD5870 27 fps

Unigine tesselation, 8x AA

RadeonHD5850 19 fps
GeForceGTX470 20 fps
RadeonHD5870 23 fps
http://www.realworldtech.com/forums/?action=detail&id=107906&threadid=107906&roomid=2

Hope Fermi is better than this
 

If that really is the final version of the 470 then as it's faster than the 5850 it stands to reason that the 480 will be quicker than the 5870, fastest single GPU card honours once more go to Nvidia.
 


300 watts 480 beating 190 watts 5870, 6 months later? Wow what an accolade. :na:
 
300 watts 480 beating 190 watts 5870, 6 months later? Wow what an accolade. :na:
Click to expand...
I think 190w is a bit on the low side do you have quantifiable proof or did you just pull that number out of the air like the 300w figure seeing as how the card has not been released yet, besides faster is just faster. :kaola:
 

Then don't buy one if it's not good enough for you, it's not like anyone's forcing you to.
 
188 was the rated TDP of the 5870, according to AMD, but that number seems consistent with multiple review sites. I realize they measure system power draw at the socket, but you can infer the number is at least fairly accurate based on teh draw of the other cards in an identical system.

I know you love nVidia, but that number certainly wasn't pulled out of any asses.
 
Its not whether I want one, its whether enough people will want one, given the perf, unless the price will astound.
Tho I havnt seen great [ricing either, but who knows.
nVidia needs these cards to smack ATI hard, this is more a kiss than a smack
 


188w max 5870
http://en.wikipedia.org/wiki/Comparison_of_ATI_graphics_processing_units#Evergreen_.28HD_5xxx.29_series

300w 480
http://www.fudzilla.com/content/view/17921/1/

And semiaccurate and others say 280w-300w. Don't forget this is also a 50% larger chip so "winning" by a few percent isn't impressive at all.
 
Status
Not open for further replies.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom