The GT300/Fermi Thread - Part 2!

[natty]

It's a failure only if it doesn't sell well, I doubt that.

they better price it at 200/300 then

[kadozer]

It's a failure only if it doesn't sell well, I doubt that.

This would be putting the cart before the horse. This is assuming they can even get enough units out there after initial paper launch.

[insurgent]

they better price it at 200/300 then

I doubt that too

This would be putting the cart before the horse. This is assuming they can even get enough units out there after initial paper launch.

Yeah I know, I just don't think it will be that bad, maybe wishful thinking, I just want prices to go down

[RaZz!]

rofl, this pic is so full of win

Zotac beat them to it.

this one as well. hehee.

@annihilat0r: it's in the 2nd post in this thread, where all kinds of information is gathered.

[annihilat0r]

When I had put that pic here my post was deleted. ??

[Thraxis]

I know it's FUD but figured it is relevant http://www.fudzilla.com/content/view/18072/1/

It is quite natural that the most famous and yet delayed chip called Fermi won’t be cheap but it turns out that it won't be that expensive either.


People have expected single GPU Geforce GTX 480 to end up at close to $600 or €600 in Europe after tax but it turns out that the amount Nvidia wants for its new high end card should be closer to €450.

Note that high demand might easily make shops push this price even further but Nvidia's suggested retail price should end up around €450 including VAT. Higher prices will be reserved for some dual card that is supposed to come in the next few months.

Since most available Radeon 5870 cards are selling at €375 and higher, we believe that Nvidia's price is quite OK, and some partners have told us that Nvidia might have plenty of cards to launch on the 27th.

[Andrea deluxe]

€450....

good good

[Your_Boss]

Just a question, but I read somewhere that Unigine was saying its Heaven benchmark v1.1 IS the same as the 2.0. So why have a 1.1 if it's SAME as 2.0? Just seems kind of strange to me. I'm sure there are SOME differences, question is, what?

Also, everyone has seen that GTX480 Heaven bench video where it shows the v1.1 and my question is...does ATI have a version 1.1 too? If not, then hmmmm.

[Picao84]

Hello,

This is my first post here.
At B3D forums, someone posted this image:



Its interesting, but the text written in the picture doesnt make sence with that box. By squiting and zooming at it I really see:
H = 2.08
W = 2.28

Nowhere i see the 23mm, with the reached conclusion of 529mm2.
But the H*W with the picture numbers give a pretty interesting number: 474mm2, close to the other rumored 480...

[mindfury]

http://forum.beyond3d.com/showpost.p...postcount=3601

[Olivon]

We want B1 chips !!!

[mindfury]

more pics:
http://bbs.ngacn.cc/read.php?tid=318...er_by=&&page=1

[Picao84]

We want B1 chips !!!

LOL! You dont even know for sure if a B1 is needed, or if it is worth it

[saaya]

1 485/1 300MHz is the max on HWBot, on Vantage :

NickShih's 3DMark Vantage - Performance score 31007 marks

hooooly sht! thats impressive!

[saaya]

http://forum.beyond3d.com/showpost.p...postcount=3601

http://img.ngacn.cc/attachments/mon_...df9952cd99.jpg
http://img.ngacn.cc/attachments/mon_...df3ba28b32.jpg
http://img.ngacn.cc/attachments/mon_...debabf411a.jpg

hmmmmm looks like thats a bunch of 480 chips that clocks to 800mhz+?
i wonder if thats on air though? i doubt it tbh, especially after seing those weird water marks on the ihs so my guess is that this is on water or even chilled water, which is very common for evaluation of chips.
800-900 on chilled water... then my guess is those chips will do ~650/700 on good air... nothing really new then...

[Picao84]

hmmmmm looks like thats a bunch of 480 chips that clocks to 800mhz+?
i wonder if thats on air though? i doubt it tbh, especially after seing those weird water marks on the ihs so my guess is that this is on water or even chilled water, which is very common for evaluation of chips.
800-900 on chilled water... then my guess is those chips will do ~650/700 on good air... nothing really new then...

I would like to know from where are you taking the conclusions about clocks?

[saaya]

oh i didnt see that they are numbered linearily... i thought i saw several with the same number

[Olivon]

LOL! You dont even know for sure if a B1 is needed, or if it is worth it

Yeah, das' right, I listen rumours too much. Not very clever from my side.

[saaya]

well, its VERY unlikely that B1 is going to make things worse, and its very likely that B1 will improve clockspeeds and reduce power, which means higher clocks as well as heat isnt limiting that much anymore...so it does make sense to ALWAYS want a new stepping

[Picao84]

well, its VERY unlikely that B1 is going to make things worse, and its very likely that B1 will improve clockspeeds and reduce power, which means higher clocks as well as heat isnt limiting that much anymore...so it does make sense to ALWAYS want a new stepping

Well if Charlie is right and the architecture is broken (playing devils advocate here :p), would a B1 improve things that much?

[annihilat0r]

AFAIK every stepping pushes production 2 months or so (charlie says). It must be the case that Nvidia currently is in the process of designing B1, or it has already been put in the oven. Which probably means that 2 or 3 months after launch we can see the B1 stepping - probably with a full GF100 with 512 SP's, and the dual GTX 470. God, this made so much sense. I think I'll apply for a CEO position at Nvidia.

[annihilat0r]

Well if Charlie is right and the architecture is broken (playing devils advocate here :p), would a B1 improve things that much?

What does Charlie know about the architecture that he can say it's the architecture that is botched and causing all the problems ? Maybe it just needs more time to be designed efficiently. If Nvidia engineers made a decision to go some way, well it can be the wrong way to go, but it can't be so bad as to not allow for the development of a decent graphics chip.

After all he had put the GF100's at the WORST possible light, saying that they are broken and completely unmanufacturable, but here is Nvidia (or TSMC), actually manufacturing the chips. Nothing can be THAT bad.

[Zed_X]

annihilat0r ... yeah man, you are almost right! The biggest launch in NV history is planned to this summer.

[Picao84]

AFAIK every stepping pushes production 2 months or so (charlie says). It must be the case that Nvidia currently is in the process of designing B1, or it has already been put in the oven. Which probably means that 2 or 3 months after launch we can see the B1 stepping - probably with a full GF100 with 512 SP's, and the dual GTX 470. God, this made so much sense. I think I'll apply for a CEO position at Nvidia.

Hmm.. I dont think nVIDIA would trust you.. Common you are basing your conclusions on the guy who wants more to see nVIDIA fail

[Chickenfeed]

Given it is most likely A3 will be the shipping part, any benifits of B1 will only effect adopters later on down the road so unless it offers some tangible overclocking headroom, I'd say its nothing to be excited about from a consumer standpoint ( given Nvidia will ship with their reference clocks, improvements in yields and steppings won't change this down... ; ie the 480 will have the same clocks from birth to death )

I still highly doubt the viability of a dual gpu GF100 on 40nm... it would have to be so cut down its not even funny.