Nvidia Pascal Thread

Megaman

Megaman

Well-known member
What is being speculated.

Codename GP100 for enthusiasts

Codename GP104 for high end

Codename GP106 for Performance

Codename GP107 for mainstream

NVlink support
HBM 2
Most likely 16nm

Die size speculated to be 500-550mm²

If HBM 2 then more than 4GB VRAM will be an option.

Word on the street is HBM 1 will be skipped and HBM 2 will be the first of the new memory designs to start off.

ETA Q3 2016. Most of the info is speculation right now but more info will be released in the next couple of months.

YOG5ok8.png


http://www.3dcenter.org/news/nvidias-gp100-chip-der-pascal-generation-hat-angeblich-seinen-tape-out-absolviert

http://www.3dcenter.org/news/wie-nvidias-pascal-generation-fuer-das-jahr-2016-aussehen-koennte
 
Last edited:
I was hoping for pascal in q1 for my Oculus :cry:

Well then.. hope the prices for 980 Ti comes down then and I will get that instead
 
I am eagerly awaiting the next round of HBM products.

My 690 is still plugging along swimmingly, since I am playing at 1200p, but with bigger, better, higher-res monitors, expectations changing a touch :)
 
Megaman said:
It seems WCCF is very confidant about this.





http://wccftech.com/nvidias-gp100-pascal-flagship-pack-4096-bit-memory-bus-8hi-hbm-stacks/
Click to expand...

and they repeat

Also unlike Nvidia, AMD has a much more powerful incentive to launch its next generation of FinFET GPUs first. This is because the company has priority to HBM2 capacity – which is going to be limited initially – as a result of co-inventing the technology with Hynix. By pushing its graphics products to launch first AMD can establish two competitive advantages over its rival.
Click to expand...

I still don't think we will see either before the end of next year
 
Yes, we know AMD has priority on HBM 2.

I do hope AMD can settle in with HBM 2 for quiet some time before NV can release theirs.
 
Megaman said:
Yes, we know AMD has priority on HBM 2.

I do hope AMD can settle in with HBM 2 for quiet some time before NV can release theirs.
Click to expand...
we might see a 8gb fury x with hbm 2.0 but not till q2 next year
but before we see fury 2 and ( 1080 ) small Pascal
 
bill dennison said:
we might see a 8gb fury x with hbm 2.0 but not till q2 next year
but before we see fury 2 and ( 1080 ) small Pascal
Click to expand...

I can't see Nvidia sitting on their hands while AMD launches it's Greenland GPU. I think NV will definitely launch a 4gb HBM1 Pascal card to try and pee on AMD's new cards. If they can launch a 16nm HBM1 card that should be able to beat out the first gen Fury's and give people time to pause before thinking about buying an AMD HBM2 card.

For AMD to capitalise on their first mover advantage they will have to do several things they have failed abysmally at in the past. First have a strategy and leadership that delivers. Secondly execution which has always plagued them and thirdly get the marketing behind the cards at launch and make sure there are lots of cards in the channel to sell at launch.

TBH if you look at Fiji they failed on all three!! I'm therefore not so confident they will be able to deliver next year either and each time they fail to deliver the window of opportunity gets smaller. Hopefully they can surprise everyone but I wouldn't want to put my money on it.
 
I think at this point we need everybody on the new smaller process nodes as soon as possible. Stuff is starting to stagnate at 28nm. I think NV could squeeze a little more out of it but I have my doubts that AMD can.
 
bittermann said:
I think at this point we need everybody on the new smaller process nodes as soon as possible. Stuff is starting to stagnate at 28nm. I think NV could squeeze a little more out of it but I have my doubts that AMD can.
Click to expand...

Didn't SusieSu say that a Fury X2 was inbound this year? If so, that'll most likely be 28nm.
 
bittermann said:
I think at this point we need everybody on the new smaller process nodes as soon as possible. Stuff is starting to stagnate at 28nm. I think NV could squeeze a little more out of it but I have my doubts that AMD can.
Click to expand...

Yup, we've pretty much reach the upper limit of this process with GPU's.
 
Seems like Pascal will be one beast of a card, especially NV's with use of memory optimizations. Hope that transfers to HBM as well.
 
Last edited:
You must log in or register to reply here.
Back
Top