[ Various] NVIDIA could launch Pascal-based mobile GPUs by end-2016

[KaRLiToS]

http://vr-zone.com/articles/nvidia-could-launch-pascal-based-mobile-gpus-by-end-of-2016/104874.html
http://www.tweaktown.com/news/49815/nvidia-launch-next-gen-pascal-gpus-hbm2-2h-2016/index.html
http://www.gulli.com/news/27035-neue-details-zu-nvidias-pascal-chip-2016-01-22

Recently a rumor claimed NVIDIA was working on the GTX 970MX and 980MX GPUs for notebooks. However, a new report from Notebookcheck claims that is not the case.

The website claims to have heard from various industry insiders that NVIDIA has no plans of updating the GTX 970M and GTX 980M cards, but will instead focus on new mobile GPUs based on the next-gen Pascal architecture. The first Pascal based mobile GPUs are expected to be released either by the end of this year or by early next year.

Hmmm, if big die Pascal comes first, I might pull the trigger.

 

[NTMBK]

If they're only expecting mobile Pascal by end of 2016/early 2017, sounds like they're ~6 months behind AMD. Hope they're out sooner than that

 

[KaRLiToS]

In the tweak town article they say 2H 2016, it could be From July to December 2016 hopefully

Now that HBM2 is beginning to flow into the market, thanks to Samsung making 4GB HBM2-based DRAM, NVIDIA is getting confident with Pascal - with the latest rumor stating that the company will unveil its next-gen GPUs in the first half of this year, with availability to follow in 2H 2016.

 

[thilanliyan]

Does nVidia usually release mobile or desktop gpus first?

 

[ShintaiDK]

Desktop.

Mobile tends to be a month or so after.

 

[Techhog]

So, Polaris hits desktop and laptop in June, Pascal hits desktop in June and laptop in July?

 

[Stuka87]

So, Polaris hits desktop and laptop in June, Pascal hits desktop in June and laptop in July?

Its not clear when nVidia will release. They didn't say Q3, they said 2H, so it means they are not sure on an exact time yet either. Going off the last time there was a node change, they were ~5 months after AMD.

 

[ShintaiDK]

Its not clear when nVidia will release. They didn't say Q3, they said 2H, so it means they are not sure on an exact time yet either. Going off the last time there was a node change, they were ~5 months after AMD.

You mean 3 months? 680 was march, 7970 was January.

 

[tviceman]

Nobody knows ANYTHING. All of this is based on what a tech writer wrote based on his on assumptions. Here's what we do know: Drive PX2 will use Pascal GPU's and I'm pretty sure JHH said during CES would be available in Q3 of this year. So that means there will be at least 1 Pascal chip in full production by late Q2 or early Q3 of this year.

VR-ZONE says: "Key rival AMD is also developing new mobile GPUs based on its Polaris architecture, but those too aren’t likely to be launched this year, so NVIDIA doesn’t really have to worry about any competition in the near future" which is crap, so case in point nobody knows.

 

[littleg]

If early 2017 is being mentioned as a possibility then the earliest i'd expect would be Q4 2016.

 

[Samwell]

Nobody knows ANYTHING. All of this is based on what a tech writer wrote based on his on assumptions. Here's what we do know: Drive PX2 will use Pascal GPU's and I'm pretty sure JHH said during CES would be available in Q3 of this year. So that means there will be at least 1 Pascal chip in full production by late Q2 or early Q3 of this year.

Q2 for sampling, Q4 availability was mentioned as far as i know. So taking into account that PX2 might need more evaluation time than gpus end of q3 or early q4 this gpu should pop up. But this is just GP106. What's with GP104, GP107? No one knows. So like you say, nobody knows anything, despite that it seems no 980/970MX is coming. But at least we know that they have at least one Pascal chip running in lab since a few months and as this thing has probably an interposer of 900mm² size it has to be GP100. So we should see leaks in the next months.

 

[tviceman]

Q2 for sampling, Q4 availability was mentioned as far as i know. So taking into account that PX2 might need more evaluation time than gpus end of q3 or early q4 this gpu should pop up. But this is just GP106. What's with GP104, GP107? No one knows. So like you say, nobody knows anything, despite that it seems no 980/970MX is coming. But at least we know that they have at least one Pascal chip running in lab since a few months and as this thing has probably an interposer of 900mm² size it has to be GP100. So we should see leaks in the next months.

Keep in mind that with both Kepler and Maxwell the X06 chip was the third chip released to market with their GPU families, behind both the low end X07 and higher end X04 chip. If PX 2 is based off the GP106 chip, chances are GP107 and GP104 are in the labs running too. But again, nobody actually knows anything for sure. When was the last time Nvidia demoed an unreleased, undisclosed GPU to the public? They certainly never did with either Maxwell or Kepler. The fact that they haven't shown or talked about anything means literally nothing.

 

[Vesku]

If Nvidia is even the "normal" 3 months behind that may leave the entire back-to-school and most of the holiday season to AMD. I always thought the aggressive GTX 970 pricing was a way to get as many units out as possible before AMD FinFet.

 

[Silverforce11]

When was the last time JHH had to present a fake card to the press & investors?

Wood-screw Fermi = delays.

Fake "980M" Pascal = delays.

New node, new tech. Delays = expected. Yes, even Intel got delayed, but it didn't matter to them since they had no competitors..

If AMD can actually execute Polaris( & Zen) without delays (ahead of their competitors), that is the exception.

 

[Arachnotronic]

When was the last time JHH had to present a fake card to the press & investors?

Wood-screw Fermi = delays.

Fake "980M" Pascal = delays.

New node, new tech. Delays = expected. Yes, even Intel got delayed, but it didn't matter to them since they had no competitors..

If AMD can actually execute Polaris( & Zen) without delays (ahead of their competitors), that is the exception.

AMD has no chance of executing Zen "ahead of [its] competitors." Skylake is here today and Broadwell-E is coming soon.

 

[MrTeal]

Q2 for sampling, Q4 availability was mentioned as far as i know. So taking into account that PX2 might need more evaluation time than gpus end of q3 or early q4 this gpu should pop up. But this is just GP106. What's with GP104, GP107? No one knows. So like you say, nobody knows anything, despite that it seems no 980/970MX is coming. But at least we know that they have at least one Pascal chip running in lab since a few months and as this thing has probably an interposer of 900mm² size it has to be GP100. So we should see leaks in the next months.

Wait, where did you see nVidia was running a Pascal HBM chip on an interposer in the lab for the last few months?

 

[Silverforce11]

AMD has no chance of executing Zen "ahead of [its] competitors." Skylake is here today and Broadwell-E is coming soon.

That's like it was in (), because the focus here is Polaris. Zen is something AMD needs to execute to be competitive because we all know how far they are behind against Intel.

 

[tviceman]

When was the last time JHH had to present a fake card to the press & investors?

Wood-screw Fermi = delays.

Fake "980M" Pascal = delays.

New node, new tech. Delays = expected. Yes, even Intel got delayed, but it didn't matter to them since they had no competitors..

If AMD can actually execute Polaris( & Zen) without delays (ahead of their competitors), that is the exception.

You are reading way, way, way too much into this. Not only do I remain correct in what I said - nobody knows squat - but you're basically (as always) cheer leading and speaking from the heart. You're assuming and hoping AMD executes perfectly assuming and hoping Nvidia falls on it's face. You expect Nvidia to be delayed and having issues because they once had delays and issues SIX YEARS AGO. ATI once had delays and issues too, but no one keeps bringing that up...

AMD often shows off pre-production silicon in some way. They either show it running a game without giving away frame rates or they have sponsored developers hype it up via twitter. Nvidia *never* shows off pre-production silicon running. They never have their sponsors talk about unannounced hardware either. Everything that is happening right now on both sides is par for the course.

 

[Silverforce11]

Note I said IF AMD can execute without delays...

That would be the exception to the rule: New unproven node, new uarch, potentially new memory tech.. how do you not expect there will be some delays?

Really, if they can pull it off, I will be very surprised given their financial & resulting R&D funding vs the competition.

 

[Arachnotronic]

Note I said IF AMD can execute without delays...

That would be the exception to the rule: New unproven node, new uarch, potentially new memory tech.. how do you not expect there will be some delays?

Really, if they can pull it off, I will be very surprised given their financial & resulting R&D funding vs the competition.

Right on, IF they can pull it off with Polaris (beating NV to the punch by a material amount of time), I will be VERY impressed & will become an AMD GPU customer again.

As I have said on these forums, the first co to put out a high end 14/16nm GPU gets my money. I don't care if that is AMD or NV.

 

[swilli89]

I think we are all in for a possible let down when all we see are ~250mm^2 products for the majority of the year. I think what it really boils down to for both AMD and NV's respective time tables are:

TSMC brand new 16nm node and feature size versus Samsung revised 14nm node. Samsung probably takes the edge here, they have been in high volume production of 14nm for almost a full year now.

What I truly have no idea about is if AMD will be helped by already having experience develop and produce products with HBM memory.

The above two things are probably the only way AMD would out-execute NV with its higher R&D budget.

As others have said, I'll probably take the first high end product that I can get.

 

[JDG1980]

If it really takes Nvidia until near the end of the year to get Pascal-based mobile GPUs to market, this could be very good news for AMD. We know AMD already has the small Polaris 10 chip demo-ready, and this chip seems tailor made for mobile applications. It could potentially hit the market by Q2, especially if AMD is able to use Samsung for yields until GloFo gets their copied process up to speed. We can already pencil in Polaris 10 on the next 15" MacBook Pro (which currently is stuck on Cape Verde), but there are probably other design wins to be had as well. Polaris 10 is clearly going to provide better performance and perf/watt than GM107, so if Nvidia can't come up with GP107 until the end of the year, they could lose out on holiday season sales of mid-range gaming laptops.

I'm not expecting to see massive profits by any means (AMD has too many fixed legacy costs based on Hector Ruiz's bad decisions), but Polaris might be able to stop or at least slow the financial bleeding until Zen is ready for release in 10-11 months.

 

[littleg]

Yep, big opportunity for AMD. Just a case of waiting to see how they manage to screw it up this time.

 

[Samwell]

Wait, where did you see nVidia was running a Pascal HBM chip on an interposer in the lab for the last few months?

At supercomputing 2015 in november it was mentioned in talks of nv engenieers and i believe this guys way more than any press folks. It seems the Zauba information is right and they have GP100 silicon since July 2015. So GP100 will launch first probably in a few months, of course just in HPC, not consumer. But other gpus might be late because of NVs concentration on the big chip.

They don't have to demo it in public like AMD because the circumstances are different. AMD wants publicity so people wait for polaris before buying a gpu now. Nv is selling extremely well at the moment, so no need to danger this by showing something of pascal.

 

[ShintaiDK]

They don't have to demo it in public like AMD because the circumstances are different. AMD wants publicity so people wait for polaris before buying a gpu now. Nv is selling extremely well at the moment, so no need to danger this by showing something of pascal.

Exactly. We see the same with CPUs. AMD is busy for ages now talking about Zen as well. While Intel...not so much.