Nvidia Announces PCIe version Tesla P100
Click here to post a comment for Nvidia Announces PCIe version Tesla P100 on our message forum
harkinsteven
Check spelling^
Denial
I wish they had a vendor list of companies that are buying this, for my own curiosity. I know Google has their own Tensor Processing Unit which they claim is the best in the industry for deep learning. I wonder how these stack up.
Cave Waverider
I assume we could see something similar as the PCIe versions for the Geforce Ti and Titan Pascal variants, respectively.
BD2015
schmidtbag
Denial
XenthorX
I might be flamed but... Can you play games with this card?
I know about all the deep-learning focus, but "what if" ?
schmidtbag
XenthorX
Denial
It's also passively cooled by the server fans. So unless you have some wind tunnel case cooling going on, you're going to have a bad time. Then there is also driver support and stuff. Oh and it probably costs over $10,000+ per card.
You probably can but it would require some dedication.
Koniakki
fantaskarsef
So my question is, and I know I'm hitting a nerve here probably with asking, but are the Titans and 1080TIs as we expect them going to be based on that GP100 chip, or are we getting some kind of cut down GP102 chip, missing the double precision part and nothing else, as in full single precision and memory bandwidth and everything?
Denial
Koniakki
WareTernal
http://www.top500.org/
https://www.youtube.com/watch?v=J6oE5Knk8K8
I don't think they will give you a list, but they claim they are selling a bunch to the U.S. Department of Energy. In 6 months, you can probably look at the top500 and see if anyone is using them.
rl66
rl66
Spets
So we might see a Titan or GeForce version this year? 😀
moab600
I think there is a chance that nvidia will counter Vega with a cut down titan version, if they can't produce enough units for the full version.
Or they release it in Feb-March.
Aelders
Gp100 isn't coming to GeForce guys, they're not going to waste all that die area on fp64, not to mention that the size of SMs are cut in half, so register file size is doubled across the chip. HBM won't be worth the cost.
There's probably going to be another gpu called gp102 for GeForce and quadro variants with 12/24 gb g5x