Nvidia PASCAL...next-gen GPU.

[the54thvoid]

Simple answer... The 1st to implement HBM was AMD.

Nvidia will take this tec and Re-engineer the HBM shit and stomp AMD as usual for this go Around But----

But yes I say, as being a red Team fan.. umm that the Green team will prevail as the usual k

I dunno, AMD have experience at working hardware level of HBM, Nvidia haven't. There's a good chance AMD can utilise HBM2 more effectively given their hands on experience with Fiji. After all, its not just HBM that's immature, its the design implementation of it. Surely AMD have a head start?

 

[fullinfusion]

I dunno, AMD have experience at working hardware level of HBM, Nvidia haven't. There's a good chance AMD can utilise HBM2 more effectively given their hands on experience with Fiji. After all, its not just HBM that's immature, its the design implementation of it. Surely AMD have a head start?

Agreed @the54thvoid

Amd Have there hand in the honey pot as they'd say....

but being A red team supporter.................. I really don't know, nvidia.. always seem to overcome the red team...

I never give up, but I always take the reality of these new cards to what they are and not what they do.. Umm lol, Yeah the both Is what I look at Duhh

 

[arbiter]

From what I see, Pascel is just NVIDIA's interpretation of AMD Fury: die shrink, some tweaks, HBM2 controller.

It was good that nvidia didn't use HBM on 980ti, else the shortage of fury's you see now would be a lot worse otherwise.
I could also see Nvidia getting attacked over 4gb ram where as AMD gets a pass at it which they have.

Did AMD actually invent HBM? Or are they not just the first to utilise it?

They didn't invent it, they worked with Hynix to make it so if anyone would own the patent for it would be Hynix.

 

[Xzibit]

They didn't invent it, they worked with Hynix to make it so if anyone would own the patent for it would be Hynix.

AMD and Hynix announce joint development of HBM memory stacks

AMD credits Brian Black for solving the stacking.

 

[Silas Woodruff]

I dunno, AMD have experience at working hardware level of HBM, Nvidia haven't. There's a good chance AMD can utilise HBM2 more effectively given their hands on experience with Fiji. After all, its not just HBM that's immature, its the design implementation of it. Surely AMD have a head start?

Don't think that really matters too much, Nvidia can learn from AMD HBM 1 mistakes to make HBM 2 better.

 

[arbiter]

Don't think that really matters too much, Nvidia can learn from AMD HBM 1 mistakes to make HBM 2 better.

Also nvidia has prototypes of their next gpu with HBM2 which was reported like 1-2 months ago so they have time to work with HBM2.

 

[Xzibit]

Don't think that really matters too much, Nvidia can learn from AMD HBM 1 mistakes to make HBM 2 better.

Its just memory and who ever utilizes it better through their GPU architecture will lead.

 

[BiggieShady]

NVidia has very good memory controller engineering team that did a great job since Kepler, I kinda expect them to pull the proverbial rabbit out of the hat.
Regarding NVLink, remember when motherboards had northbridge chip on them, those times when chipset was in fact a set of chips ... I don't think that motherboard manufacturers would mind putting one more proprietary chip while charging a fat premium ... I'm talking gsync kind of premium ... it'll work beautifully and cost a fortune ... in onther words, it will suck

 

[terroralpha]

Amd made this and all Nvidia did was copy this!

well, no. AMD came up with the concept for it, but Hynix "made it". whatever the case may be, stacked RAM was coming anyway. if Hynix would not have done it, Micron, Samsung or even Intel would have. they all have working 3D volatile and non volatile memory technology.

while AMD obviously deserves a lot of credit for conceiving the idea, they had no choice but to make this technology open to everyone, even the dirty bastards at nvidia. no memory manufacturer would ever drop a massive fortune developing something specifically for one company who happen to be in distant second place in every industry they are in, and those industries are slowly dying. Hynix survives on small margins and insane volumes.

 

[cadaveca]

well, no. AMD came up with the concept for it, but Hynix "made it". whatever the case may be, stacked RAM was coming anyway. if Hynix would not have done it, Micron, Samsung or even Intel would have. they all have working 3D volatile and non volatile memory technology.

while AMD obviously deserves a lot of credit for conceiving the idea, they had no choice but to make this technology open to everyone, even the dirty bastards at nvidia. no memory manufacturer would ever drop a massive fortune developing something specifically for one company who happen to be in distant second place in every industry they are in, and those industries are slowly dying. Hynix survives on small margins and insane volumes.

It makes sense that AMD would have much experience with TSV design, since they had issues wit ha certain GPU that on it's refresh, got doubled-up TSVs (RV740... LINK )... Which makes all of this rather... well... old tech, just not public tech.

David Wang, ATI’s VP of Graphics Engineering at the time, had concerns about TSMC’s 40nm process that he voiced to Carrell early on in the RV740 design process. David was worried that the metal handling in the fabrication process might lead to via quality issues. Vias are tiny connections between the different metal layers on a chip, and the thinking was that the via failure rate at 40nm was high enough to impact the yield of the process. Even if the vias wouldn’t fail completely, the quality of the via would degrade the signal going through the via.

That linked article might have been one of Anandtech's best articles ever...

 

[Ebo]

To me Pascal isent really a giant leap forward, its more like a normal development.

Do I like what have been rumors so far, not really. To me its like Nvidia just lets others do the hard work, and prosper on others achivement.

what I really dont understand is why Hynix and AMD didnt take out a patent on this whole new ram teknology, and just kept Nvidia out of it all. The reason for it would force Nvidia to make their own kind of new tec and most positivily set back development of new GFX by at least a couple of years which I see as a good thing.

I would rather have both compagnies release new products when they really have something to show, not just the mishmash they have been releasing the last years.

 

[arbiter]

To me Pascal isent really a giant leap forward, its more like a normal development.

Do I like what have been rumors so far, not really. To me its like Nvidia just lets others do the hard work, and prosper on others achivement.

what I really dont understand is why Hynix and AMD didnt take out a patent on this whole new ram teknology, and just kept Nvidia out of it all. The reason for it would force Nvidia to make their own kind of new tec and most positivily set back development of new GFX by at least a couple of years which I see as a good thing.

Pretty sure hynix if they had patent for it would be very happy to sell to nvidia cause well it will bring in a ton of money due to nvidia having the market. If they did that nvidia likely would come up with their own ram in partnership with one hynix's competitors. Hynix will only have AMD to sell to which is big question if AMD will survive. As for you saying they should shut nvidia out, well that would be very hypocritical of AMD to do that to claim they want all open standards this and that and end up shutting out nvidia from it. Early rumored reports is pascal will be a massive jump in performance think remember seeing along likes of like 50-60%. Nvidia cards as fast as they are tend to be more bottlenecked by memory bandwidth since kepler series was released. That memory would remove any of that.