[Kitguru]Nvidia`s big Pascal GP100 have taped out - Q1 2016 release

[ShintaiDK]

Kitguru is merely the intermediary.

Kitguru writes:

"An anonymous person presumably with access to confidential information in the semiconductor industry revealed in a post over at Beyond3D forums that Nvidia had already taped out its next-generation code-named GP100 graphics processing unit. Nowadays, a tape-out means that the design of an integrated circuit has been finalized, but the first actual chips materialize only months after their tape-out."

Did you even read the message?

An anonymous person
presumably with access to confidential information


Really guys, this is remarkable. How much leeway are they giving this source.

I think I should start creating some leaks myself.

So whats your edvidence that nVidia is behind? Fake news and articles shouldnt be new to you with all the Fury information for example.

 

[Phynaz]

They should had better not tempted such a scam as it s aknowledgment that they are in full panic mode..

http://semiaccurate.com/forums/showpost.php?p=237272&postcount=1120

100% accurate and reliable.

It's called Semiaccurate for a reason.

From your linked post:

Everything I post in these or any other forums, or anywhere else online for that matter, is strictly my personal opinion.

 

[ShintaiDK]

I've seriously considered this. I wonder how hard it would be to get WCCFTech to report my speculations about the next Mac Pro as fact...

Its uncanny easy. I tried it with a few frinds as a practical joke. A lot of sites pretty much eat up anything you serve without any form of source or material critisism.

 

[Cloudfire777]

What a complete waste of time this forum is.

 

[Phynaz]

Why, because people don't agree with you?

 

[maddie]

Its uncanny easy. I tried it with a few frinds as a practical joke. A lot of sites pretty much eat up anything you serve without any form of source or material critisism.

What a way to discredit the OP.

My hero.

 

[AnandThenMan]

What a complete waste of time this forum is.

There is a simple solution to that.

Anyway, how can people at this point trust TSMC's guidance on node roll outs? We were supposed to be on 28nm two years before it actually happened, and have been stuck on the node for what seems like an eternity. I am not blaming TSMC really the technical challenges of moving to smaller nodes has proved to be extremely difficult.

Its uncanny easy. I tried it with a few frinds as a practical joke. A lot of sites pretty much eat up anything you serve without any form of source or material critisism.

I'm glad you pointed this out, this is exactly why we don't need random and dubious sources being posted as fact.

 

[Abwx]

.
Volume production beginning in July next month

Nvidia Pascal GP100 and 16nm FinFET in early 2016 is most likely very legit

Volume production for 50-100mm2 chips eventualy, and at awfull yields, likely that a 500mm2 chip wouldnt even boot at this point, as for tape outs it is necessary to have the final specifications of the transistors, otherwise you cant tape out anything.

It s unlikely that they got the SDKs early enough to allow for a design being completed currently, the process is simply too recent, heck it doesnt exist elswhere than in TSMC slides and announcements.

All of this smell a PR campaign from a firm that know that it has nothing to counter their competitor perhaps till late 2016, never would Nvidia announce to recent customers that their cards will be obsolete in a matter of 9 months if they were sure that the GT9xx would hold the comparison, it is even counter productive since people interested by GT9XX could as well decide to wait next year releases.....

 

[Maxima1]

Why, because people don't agree with you?

Yeah, doesn't make sense. He's also the one who did that, "I've got a source that says l33t" on multiple forums...

 

[ShintaiDK]

There is a simple solution to that.

Anyway, how can people at this point trust TSMC's guidance on node roll outs? We were supposed to be on 28nm two years before it actually happened, and have been stuck on the node for what seems like an eternity. I am not blaming TSMC really the technical challenges of moving to smaller nodes has proved to be extremely difficult.

Easier with a picture or 2.

 

[AnandThenMan]

nm

 

[maddie]

Easier with a picture or 2.

So we have had 16nm Finfet since 2013?

Sounds like an early Apple maps.

I'm laughing so hard, I'm crying.

 

[Abwx]

It's called Semiaccurate for a reason.

From your linked post:

Yes but you dont know the reason apparently...

He cant state that he has factual infos without being deprived definitly of any source...

http://www.bitsandchips.it/

Why would Kitguru be more credible than Bitsandchips..?.

 

[Azix]

Not impossible this is one of the reasons why Nvidia launched GTX 980Ti so quick after Titan X. They got a new beast coming out already in February next year based on Pascal

more like this rumour and the 980ti release suggest they are doing what they can to distract from AMD

 

[PPB]

Hbm needs a rework driver wise to optimize the new memory subsystem

Big dies have the lowest possible yields at any given process, specially on ones ramping up new processes.

Tsmc has not much experience on finfet nodes, they also have a history of being late and having worse technical specs than their competitor's proceses.

Nvidia has a history of not meeting schedules when both a new architecture, a new process and a new memory subsystem are involved. They also dont have a history of good yield ramping up with their obsenely big dies as their top end skus.

Hbm1 ia still on its first steps and there is no guarantee hbm2 will be ready in meaningful quantities by the scheduled date this arcticle implies.

Anyone in their sane mind will see that article and piss their paints off laughing as it is remarkably unrealistic. Fact is you can tape out anything you want, to have a working product in the end with acceptable yields and good inventory is another matter entirely. This is just nvidia trying to steal the thunder and osbourning their current product stack in the process.

 

[yacoub]

Not impossible this is one of the reasons why Nvidia launched GTX 980Ti so quick after Titan X. They got a new beast coming out already in February next year based on Pascal

If it does drop in Feb, it'll be for industrial/commercial products not consumer-oriented gaming products.

 

[ph2000]

--ignore wrong post

 

[Pinstripe]

There's always a chance they drop the small GP106 first in Q1, kinda like GTX 750/Ti, with the full Pascal (GP200 series) following in Fall 2016. Or not.

 

[RoarTiger]

kitguru fail. Hynix reported HBM2 availability in Q2, but now somehow Pascal is Q1. Either Pascal wont be HBM2 (which would be actual news and should be the title of the article) or this is 100% clickbait garbage.

http://www.kitguru.net/components/g...mory-chips-opens-way-for-32gb-graphics-cards/

 

[stuff_me_good]

So nvidia has gone for panic mode... what a surprise.

TSMC roadmaps are joke, nothing more that smoke and mirrors again.

 

[JDG1980]

There's always a chance they drop the small GP106 first in Q1, kinda like GTX 750/Ti, with the full Pascal (GP200 series) following in Fall 2016. Or not.

Wouldn't that be GP107, not GP106? The usual Nvidia layout is 8->7->6->4->0 smallest to biggest. GTX 750/Ti was GM107.

 

[happy medium]

I see you ignored my post

Get used to it, by now most of us do. Your posts are more of the same, Nvidia sucks , AMD will do it better. With a bunch of techno bullsh!t in the middle.

 

[Alatar]

Someone please explain to me why people always consider rumors they don't like the work of paid shills? I don't quite get it.

Also why would anyone making an "nvidia shill" claim link to a semiaccurate post? SA and SA forums are basically the modern version of AMD zone.

 

[maddie]

Get used to it, by now most of us do. Your posts are more of the same, Nvidia sucks , AMD will do it better. With a bunch of techno bullsh!t in the middle.

I'm glad you read it. Keep it up.

 

[Abwx]

Someone please explain to me why people always consider rumors they don't like the work of paid shills? I don't quite get it.

So they are not automatically shills, quite possible after all, but then :

Also why would anyone making an "nvidia shill" claim link to a semiaccurate post? SA and SA forums are basically the modern version of AMD zone.

So in this site they, including me, are all shills, a big thank you to help us differentiate between the good and the bad, your arguments are indeed indiscutable...