It seems Deep Learning systems are desperately needed by many here.
Just having some laughs folks, don't get angry.
It seems Deep Learning systems are desperately needed by many here.It was stated by the Nvidia CEO himself the GP100 will only be available as part of DGX-1 system. People wanting GP100 based cards for anything else will be required to wait until next year it appears. That is what I got from watching it and from what the technical press have seemed to have reported.
Calm down before you have a stroke!
Those who use the term ADF love JHH....They'd probably have his babies if they were capable.
ADF is the term they use to make fun of those who support AMD. It's the best they could come up with I guess.
It seems Deep Learning systems are desperately needed by many here.
Just having some laughs folks, don't get angry.
Calm down before you have a stroke!
Those who use the term ADF love JHH....They'd probably have his babies if they were capable.
ADF is the term they use to make fun of those who support AMD. It's the best they could come up with I guess.
Who is ADF?? The Nvidia CEO said the following during the GP100 unveil:
http://www.tomshardware.co.uk/nvidia-pascal-tesla-p100-gpu,news-52735.html
You did actually watch the event right?? It was said during that event the GP100 would ship to HPC first and be available to OEMs by Q1 2017. Even the price of the system,ie,$129000 was mentioned during that event.
It was the same with the GK110 - it first shipped to ORNL months before other customers got it.
Is ADF another name for Jen-Hsun Huang?? That would JHH right??
Edit to post!
Here is arstechnica reporting on it:
http://arstechnica.co.uk/gadgets/2016/04/nvidia-tesla-p100-pascal-details/
http://www.enterprisetech.com/2016/04/06/nvidia-unleashes-monster-pascal-gpu-card-gtc16/
http://techreport.com/news/29946/pascal-makes-its-debut-on-nvidia-tesla-p100-hpc-card
It was stated by the Nvidia CEO himself the GP100 will only be available as part of DGX-1 system. People wanting GP100 based cards for anything else will be required to wait until next year it appears. That is what I got from watching it and from what the technical press have seemed to have reported.
This slide is surprising to me. Per NV themselves, they estimated not long ago that 30% of their Install Base was using Maxwell, with the remaining 70% are still on previous gen architectures. This should put the idea to rest that GPU performance only matters for 1.5-2 years and that most PC gamers upgrade every 2 years. It's actually the opposite in that most PC gamers hold their NV GPUs beyond 2 years (and I would estimate even longer for AMD users given how well GCN has aged compared to Fermi and Kepler).
GP104 will be a great upgrade for those Fermi GTX470<->580 and GTX670<->780Ti users.
The projected demand for VR GPU upgrades is incredible too.
NV's desire to attack the > $300 market segment makes a lot of sense now since it seems there are millions of Fermi and Kepler users who skipped Maxwell, which means they skipped GTX950/960/970 at affordable price segments. That means if someone is using a 2010-2013 Fermi/Kepler GPU is going to be more likely to spend $300+ after not upgrading for so long and NV won't give this loyal customer the choice of cheaper GP106/107 upfront. Smart play to keep increasing ASP and pushing gross margins beyond 56%.
This may explain why some GTX480/GTX570/580/GTX670/680/770/780 owners are defending the rumored $600-650 prices for GTX1080 since well they had 3 to 6 years to save up. It also means none of these gamers would look at GP104 as a next gen upper mid-range chip either since their current GPU is so far behind in performance.
What's wrong with you, drama queen? US =/= Japan.
Begin shipping Tesla P100's first quarter of 2017, but prior to the shipment of the card alone, NVIDIA will release the Tesla P100 equipped with eight, the world's first "deep learning" supercomputer "DGX-1".
...
Official sales start in Japan also announced, in Japan, GDEP Solutions Co., Ltd., HPC Systems, Inc., from the three companies of Hitachi, Ltd., will be sold from 26 days. There was no announcement about the price, 129,000 dollars (1,419 million and $ 1 = 110 yen equivalent) in the United States to be sold in the announced at GTC 2016 's.
This may explain why some GTX480/GTX570/580/GTX670/680/770/780 owners are defending the rumored $600-650 prices for GTX1080 since well they had 3 to 6 years to save up. It also means none of these gamers would look at GP104 as a next gen upper mid-range chip either since their current GPU is so far behind in performance.
If high-performance graphics cards like Titan X, Fury X and GTX 980 Ti aren't enough to satisfy your lust for top-tier PC hardware, this year will see the arrival of new hardware with the potential to take gaming visuals and performance to the next level. Historically, AMD and Nvidia have worked hard to push PC graphics year after year, but the arrival of 14nm and 16nm chip fabrication technology using 3D FinFET transistors offers GPU vendors the first real innovation in manufacturing technology for five years. And recent data released by Nvidia suggests we're in for something really special with its upcoming Pascal architecture.
...There are plenty of other question marks we hope to see addressed soon too. For example, we know that GP100 - the 'big Pascal' chip - is designed for next-gen HBM2 memory, but what's the score with the upcoming consumer cards? Titan X and GTX 980 Ti took GDDR5 memory pretty much to its limits with a 384-bit bus paired with 7gbps modules. Will Nvidia stick with tried and tested technology, or will it go for Micron's new, higher bandwidth GDDR5X? The recent leaks of PCB shots of the GP104 see it paired with currently unidentifiable Micron chips strongly suggest that at least one of the consumer level Pascal cards will ship with the upgraded RAM [UPDATE 28/4/15 11:33am: Looks like the Micron chips have indeed been identified as GDDR5X].
...Then theres HBM2. Id have argued for its inclusion in GeForce Pascals a few months ago, but GDDR5X is on the way. This memory doubles the prefetch length and also should come with a fairly large increase in effective clock speed. Itll be cheaper to use than HBM2 at similar aggregate bandwidths, and its cheaper to implement at the on-chip PHY levelnot to mention the savings from the lack of an interposer and stack packaging. GDDR5X also doesn't have strict rules tying bandwidth to capacity. That lets Nvidia use memory sizes other than 4GB, 8GB, 12GB, or 16GB on its GeForce products, compared to the limitations of HBM2.
Given those guesses, I think theres at least one consumer chip that's still really big, but quite a bit smaller than 610 mm². It probably has similar overall throughput to GP100 in the metrics we care about for graphics, and it'll probably come with less memory capacity. Even so, it should still have plenty of overall bandwidth. Some rumours say this chip is called GP102. I think it'll have 56 to 60 SMs, 1/32nd FP64 throughput, and more than 8GB of 384-bit GDDR5X. If it exists, then its likely destined for a Titan-class card first, and maybe a enthusiasts favourite Ti product later on.
Nvidia is also likely working on a GM200 replacement for the pair of high-end GeForce non-Tis that make up the meat of the enthusiast market these days. Its likely called GP104. That chip will likely also have token FP64 throughputremember that these are GPUs, not HPC cards. I also bet it'll have 8GB of 256-bit GDDR5X, 40 SMs or thereabouts, and all the associated machinery in terms of texturing and backend throughput that implies, in a die of around 300 mm².
We’ve confirmed this information with three independent sources (Jen-Hsun, Jen-Hsun doppelganger and Elon Musk). NVIDIA Editors’ Day 2016 will take place next week in USA. It’s a special event for reviewers and technology enthusiasts who are invited for a briefing about new graphics cards. During this day NVIDIA will inform the press about new Pascal-based graphics cards.
Our sources claim that two graphics cards will be unveiled. So this should probably mean GeForce GTX 1080 and GTX 1070 showcase.
We are avoiding the term ‘announcement’ as it remains unclear if the event will be live streamed for public audience or its content will be disclosed later when NDA lifts. Normally NVIDIA tends to call public events as ‘press conferences’, so it is unlikely we will be seeing official confirmation of this Pascal event.
The last major Editors’ Day took place in 2014, a week before Maxwell GM204 announcement. So this should give you a perspective on when to we expect GeForce Pascal cards (I’m thinking maybe second or third week of May).
Two new, interesting Pascal articles today.
Eurogamer: In Theory - Can next-gen Nvidia tech offer Titan X power for GTX 970 money?
I hope they line up aftermarket models right out of the gate without a delay. Not interested in buying reference coolers anymore with how good the aftermarket stuff has gotten
Big news, Pascal Geforce(s) should be unveiled next week at NVIDIA's Editor's Day 2016:
NVIDIA Editors Day 2016
http://videocardz.com/59459/nvidia-pascal-editors-day-next-week
Even if NDA lifts later we can probably expect some credible leaks.
Why would Nvidia offer that kind of deal? They've been able to move good volume with the current pricing structure. The ~330 mm2 size dice of the P104-400 most likely costs more than the 400 mm2 GM104. 10 GHz GDDR5X RAM will cost more than 7 GHz GDDR5 RAM. So BOM costs will be up - NV will want at least the same margins (unless they have a master plan to crush AMD) - so I don't see how we get that kind of price/performance suggested by those articles - not this year.
HMM you think? I always chuckle when I see people making comments on market position and pricing strategies. You realize men much smarter and more seasoned than yourself are paid hundreds of thousands of dollars to literally make that one decision?
Why do people do this? You have literally zero insight into market strategies nvidia is and will be employing and you have no idea about their material costs. Why not leave the business analyst stuff alone and just be excited for this possibility?
HMM you think? I always chuckle when I see people making comments on market position and pricing strategies. You realize men much smarter and more seasoned than yourself are paid hundreds of thousands of dollars to literally make that one decision?
Why do people do this? You have literally zero insight into market strategies nvidia is and will be employing and you have no idea about their material costs. Why not leave the business analyst stuff alone and just be excited for this possibility?
Because it doesn't make sense. It is "theoretically" possible, but I'm not excited by fanboi theories. I've working in high tech & manufacturing as an engineer. I've figured out BOM costs. I've dealt with marketing's business analysts. Do I know for sure what NV is going to do - no, but I have a sense of what is in their best interest. Sorry if that makes you sad - put me on ignore.
I once told an Intel wireless exec that not having CDMA integrated into their cellular modems was a mistake. The exec assured me that this wouldn't be needed in the future since CDMA was on its way out.
Some time later, lo and behold, they rushed to buy the CDMA modem assets from Via Telecom.
http://www.valuewalk.com/2015/10/intel-corporation-acquires-via-telecoms-assets/
Don't assume that these people getting paid big bucks with fancy titles know everything. And don't assume that the people hanging out on this forum are just clueless business analyst wannabes -- some of us might actually know a thing or two.
I can tell you right now that only reference designs will be available at launch (many reasons for that, the most evident being to avoid leaks). However, AIB SKus will shortly follow at Computex.I hope they line up aftermarket models right out of the gate without a delay. Not interested in buying reference coolers anymore with how good the aftermarket stuff has gotten
So much for the 3 SKUs based on GP104?
Anyway, bring it on already.