More detail on the GPUs in the system:
The Nvidia DGX-1 Deep Learning Supercomputer in a Box
11–20 of 106 posts
Re: The Nvidia DGX-1 Deep Learning Supercomputer in a Box
#12Earlier quoted context omitted.
NVIDIA Tesla P100 has 21 TeraFLOPS of FP16 performance by their words. So they got 8 chips there.
Yep, they showed a diagram of how it fits together: http://i.imgur.com/xk1daFG.jpg
Re: The Nvidia DGX-1 Deep Learning Supercomputer in a Box
#13Earlier quoted context omitted.
The Nvidia slides had it at $129,000 a pop
Wow, cheap, great for startups!
As of last year prices for general HPC resources were running around $3/GFLOP[1], or about $500,000 for 170TFlops if my math is correct.
Sounds like this is a significant cost savings if it fits your use case.
Re: The Nvidia DGX-1 Deep Learning Supercomputer in a Box
#14Check out the specs here: http://images.nvidia.com/content/technologies/deep-learning/... though I'm most curious about what motherboard is in there to support NVLink and NVHS. Good overview of Pascal here: https://devblogs.nvidia.com/parallelforall/inside-pascal/ 1 question: will we see NVLink become an open standard for use in/with other coprocessors? 1 gripe: they give relative performance data as compared to a CP…
It looks like it uses a separate daughterboard that houses the GPUs + NVLink, connected to the main motherboard using quad Infiniband EDR (400Gbps) + RDMA. http://images.anandtech.com/doci/10225/SSP_85.JPG
Re: The Nvidia DGX-1 Deep Learning Supercomputer in a Box
#15Check out the specs here: http://images.nvidia.com/content/technologies/deep-learning/... though I'm most curious about what motherboard is in there to support NVLink and NVHS. Good overview of Pascal here: https://devblogs.nvidia.com/parallelforall/inside-pascal/ 1 question: will we see NVLink become an open standard for use in/with other coprocessors? 1 gripe: they give relative performance data as compared to a CP…
You mean you're not surprised that a machine with 8 GPUs, apparently costing $129k USD (from comment below), can outperform a single CPU? :)
(Of course, a better metric is that it's getting ~56x the performance at probably ~10x the TDP, but that's not surprising for a GPU with the current state of deep learning code.)
To their credit, the thermal and power engineering needed to get that dense a compute deployment is challenging. (bt, dt, have the corpses of power supplies to show for it.) But the price means that it's going to be limited to hyper-dense HPC deployments by companies that don't have the resources to engineer their own for substantially less money, such as Facebook's Big Sur design: https://code.facebook.com/posts/1687861518126048/facebook-to... . And, of course, the academics and hobbyists will continue to use consumer GPUs , which give much better performance/$ but aren't nearly as HPC-friendly.
Re: The Nvidia DGX-1 Deep Learning Supercomputer in a Box
#16Wait, how many chips did they cram in there that they're getting 170 TFlops. Even at a very generous 10 TFLOP per chip that would be 17 chips.
NVIDIA Tesla P100 has 21 TeraFLOPS of FP16 performance by their words. So they got 8 chips there.
Ah, half-floats. That explains it. Still pretty high but realistic at least.
Re: The Nvidia DGX-1 Deep Learning Supercomputer in a Box
#17Re: The Nvidia DGX-1 Deep Learning Supercomputer in a Box
#18I am looking forward to OpenCL catching up with CUDA in maturity and adoption, so that NVidia's monopoly in Silicon for deep learning will come to an end.
Re: The Nvidia DGX-1 Deep Learning Supercomputer in a Box
#19Wow, I didn't realize they were shipping HBM2 already. 720GB/s - with only 16GB of RAM, you can read it all in 22 milliseconds!
They're not. All that was mentioned in the talk was that this chip is coming soon.
Re: The Nvidia DGX-1 Deep Learning Supercomputer in a Box
#20I am looking forward to OpenCL catching up with CUDA in maturity and adoption, so that NVidia's monopoly in Silicon for deep learning will come to an end.
Me too. I really want to see some benchmarks between cuda code and opencl code generated from cuda with AMDs compiler.
Actually if anyone has a geforce/tesla get on this!