Where is Pascal? Nvidia’s Drive PX 2 prototype powered by Maxwell, not Pascal

[Mr_Troll]

Nvidia’s Drive PX 2 prototype allegedly powered by Maxwell, not Pascal

 

When Nvidia’s CEO, Jen-Hsun Huang took the stage at CES last week he unveiled the company’s next generation self driving car platform, the Drive PX 2. According to Nvidia, its Drive PX 2 platform packs the same amount of compute power as six Titan X boards, in just two GPUs. During the show, Jen-Hsun displayed the new system — but what he showed from stage almost certainly wasn’t Pascal.

The drive PX2 nvidia showed at CES:

 

 

This is what a GTX 980 MXM module ( laptop 980 ) looks like: 

 

As Anandtech reader noted, the hardware Jen-Hsun showed was nearly identical to the GTX 980 in an MXM configuration. The new Drive PX 2 is shown above, the GTX 980 MXM is shown below. The hardware isn’t just similar — the chips appear to be identical. Some readers have also claimed they can read the date code on the die as 1503A1 — which would mean the GPUs were produced in the third week of 2015.

 

If Nvidia actually used a GTX 980 MXM board for their mockup, it would explain why the Drive PX 2 looks as though it only uses GDDR5. While Nvidia could still be tapping that memory standard for its next-generation driving platform, this kind of specialized automotive system is going to be anything but cheap. We’ve said before that we expect GDDR5 and HBM to split the upcoming generation, but we expect that split in consumer hardware with relatively low amounts of GPU memory (2-4GB) and small memory busses. The Drive PX 2 platform sports four Denver CPU cores, eight Cortex-A57 CPUs, 8 TFLOPS worth of single-precision floating point, and a total power consumption of 250W. Nvidia has already said that they’ll be water-cooling the module in electric vehicles and offering a radiator block for conventional cars. Any way you slice it, this is no tiny embedded product serving as a digital entertainment front-end.Then again, it is still possible that the compute-heavy workloads the Drive PX 2 will perform don’t require HBM. It seems unlikely, but it’s possible.

 

Wood screws 2.0?

 

These issues with Pascal and the Drive PX 2 echo the Fermi “wood screw” even of 2009. Back then, Jen-Hsun held up a Fermi board that was nothing but a mock-up, proclaimed the chip was in full production, and would launch before the end of the year. In reality, NV was having major problems with GF100 and the GPU only launched in late March, 2010.The good news is, we’ve seen no sign that Nvidia is having the same types of problems that delayed Fermi’s launch and hurt the final product. As far as we know, both AMD and Nvidia are on-track to launch new architectural revisions in 2016.What’s more perplexing is why Nvidia engages in these kind of overreaches in the first place. Claiming the first public demo of a 16nm FinFET GPU may be a decent PR win, but claiming it in the context of automotive market isn’t going to ignite the coals of a GPU fanboy’s heart. Nvidia’s push into self-driving cars and deep learning is a long-term game. Launching a new Shield or game GPU might push some fans to upgrade immediately; precious few people are going to have the luxury of planning a new car purchase on the basis of Nvidia’s PX 2, no matter how powerful it might be.The entire point of holding up a product from stage is to demonstrate to the audience that the hardware actually exists. If it’s later shown that the hardware in question wasn’t what it was claimed to be, it undercuts the original point. Worse, it invites the audience to question why the company is playing fast and loose with the truth. There’s no reason to think Pascal is suffering an unusual delay — but these kind of antics invite speculation to the contrary.

 

Well done. Just realized i got fooled aswell with this drive PX2. I actually believed that those were pascal chips like Jen Sun said. 

 

Source:http://www.extremetech.com/gaming/220818-nvidias-drive-px-2-prototype-allegedly-powered-by-maxwell-not-pascal

http://www.anandtech.com/show/9903/nvidia-announces-drive-px-2-pascal-power-for-selfdriving-cars

Tomsen post in this topic  : http://linustechtips.com/main/topic/521232-nvidia-shows-off-drive-px-2-water-cooled-supercomputer-for-self-driving-cars-deep-learning-tech/

[ZetZet]

Does it matter? It's not a gaming card.

[Mr_Troll]

Does it matter? It's not a gaming card.

it does Nvidia said those were pascal GPUS.

[Guest Kloaked]

Damn. Was hoping I could finally run Crysis with my car.

[ZetZet]

it does Nvidia said those were pascal GPUS.

it's a prototype tho.

[That Norwegian Guy]

Maybe they're panicking to rework Pascal after async compute performance of Nvidia's architecture in DX12 embarrassed them?

 

Hope not, it would be great to have a whole generation of AMD recouperating market share.

[Mr_Troll]

it's a prototype tho.

but they still lied about the tech being used in it... just pointing out how they lied

[ZetZet]

but they still lied about the tech being used in it... 

why? It's a prototype, they might be using Pascal in the actual version?

 

 

Pretty sure Pascal GPU's aren't being manufactured yet, how do you want them to use Pascal?

[Sauron]

Why isn't this a gtx 990?

[Guest Kloaked]

Maybe they're panicking to rework Pascal after async compute performance of Nvidia's architecture in DX12 embarrassed them?

 

Hope not, it would be great to have a whole generation of AMD recouperating market share.

 

Don't think so. They were talking about using software (as usual) to be able to use that, and I believe they already did it with the 980 Ti IIRC and got good results compared to AMD. Can't be arsed to Google it for you though. 

[Ashaira]

the compute power of SIX TITAN X's? maybe because titan x's SUCK at compute? not that impressive.

[That Norwegian Guy]

Don't think so. They were talking about using software (as usual) to be able to use that, and I believe they already did it with the 980 Ti IIRC and got good results compared to AMD. Can't be arsed to Google it for you though. 

 

When you have a $400 card of 5000 Gflops beginning to catch up to a $600 card of 5700 Gflops due to its better async compute capabilities, (16x) it is embarrassing no matter how you look at it. I wouldn't call it "good results" if you can't proportionally outperform the 5K Gflop card with 12% more raw power.

[ivan134]

Whoa. How come the people making baseless accusations in the Polaris vs 950 thread didn't catch this?

[Guest Kloaked]

When you have a $400 card of 5000 Gflops beginning to catch up to a $600 card of 5700 Gflops due to its better async compute capabilities, (16x) it is embarrassing no matter how you look at it. I wouldn't call it "good results" if you can't proportionally outperform the 5K Gflop card with 12% more raw power.

 

If it performs well in games what does it matter?

[ivan134]

Don't think so. They were talking about using software (as usual) to be able to use that, and I believe they already did it with the 980 Ti IIRC and got good results compared to AMD. Can't be arsed to Google it for you though. 

They did catch up, but it took them months. That is the problem with Maxwell, and they're fixing that with Pascal. It saps precious resources they could be using on other things, while AMD sits back and lets their hardware do all the work. It also defeats the purpose of dx12, which is software talking directly to the hardware and less need of frequent driver releases like we're used to in dx11.

[Guest Kloaked]

They did catch up, but it took them months. That is the problem with Maxwell, and they're fixing that with Pascal. It saps precious resources they could be using on other things, while AMD sits back and lets their hardware do all the work. It also defeats the purpose of dx12, which is software talking directly to the hardware and less need of frequent driver releases like we're used to in dx11.

 

That's almost always been the case though, between the two vendors; Nvidia has almost always (least since Kepler anyways, to play it safe in my point since I'm not going to Google all of this to double check) been ahead on the software front and AMD has been chugging along in their raw performance.

 

If it works, it works. I don't need to know the details.

[Zah]

Can we shut the hell up about games, this prototype has nothing to do with games.  And as a prototype/mock-up/whatever, no one has even tested this thing, and no one has any proof the chips being used will be in the final product.

[ivan134]

That's almost always been the case though, between the two vendors; Nvidia has almost always (least since Kepler anyways, to play it safe in my point since I'm not going to Google all of this to double check) been ahead on the software front and AMD has been chugging along in their raw performance.

 

If it works, it works. I don't need to know the details.

You are right, but it will be a marketing problem for Nvidia like it is for AMD now, if people have to wait months for every dx12 release. Day 1 performance seems to be a reason a lot of people don't want AMD cards.

[Guest Kloaked]

You are right, but it will be a marketing problem for Nvidia like it is for AMD now, if people have to wait months for every dx12 release. Day 1 performance seems to be a reason a lot of people don't want AMD cards.

 

I don't think we can judge their release schedule for DX12 support on games with how long it took them to get a-sync working. Regardless of what people think about Nvidia (even AMD more so), they have intelligent people that work their asses off at the company and know what they're doing. There's no way they wouldn't be able to support it on a regular basis like they are/have been with new releases.

[GoodBytes]

You must remember that Pascal is Maxwell revision 2.0.

It has been added as Volta got delayed.

While I don't know if what was shown on stage was really Maxwell or Pascal. All I am saying that it is normal that Pascal board looks identical to Maxwell.

[Guest Kloaked]

You must remember that Pascal is Maxwell revision 2.0.

It has been added as Volta got delayed.

 

Man is that all they're doing is adding HBM and GDDR5X or whatever the shit

[don_svetlio]

Nvidia lying. So what's new?

[ivan134]

I don't think we can judge their release schedule for DX12 support on games with how long it took them to get a-sync working. Regardless of what people think about Nvidia (even AMD more so), they have intelligent people that work their asses off at the company and know what they're doing. There's no way they wouldn't be able to support it on a regular basis like they are/have been with new releases.

Ark's dx12 patch has been MIA with no sign that it's coming any time soon. It took months for them to get up to par with AMD in Ashes of the Singularity. The 980 was worse than a 390 in Fable Legends which didn't even have any indication it was using Async Compute. Can Nvidia get early working drivers in future dx 12 releases? Possible, but none of the early signs point to it happening.

[Nexxus]

but they still lied about the tech being used in it... just pointing out how they lied

They haven't lied about shit, Its not out yet.

 

If the final unit is maxwell and they haven't offered a statement correcting the pascal claims in the near future then they have lied/been dishonst.

 

Not before those conditions are met.

[samcool55]

Well, if that is pascal. I don't expect HBM.

 

So they are lying, or pascal has no HBM. Or HBM will be featured on other chips, but that would be weird because if they show it off with a big-ass headline like 6x more whatever it was.

I expected this would be at least high-end stuff and not some mid-range sh*t...