The Nvidia DGX-1 Deep Learning Supercomputer in a Box
1–10 of 106 posts
Re: The Nvidia DGX-1 Deep Learning Supercomputer in a Box
#2Any idea how much this costs?
Re: The Nvidia DGX-1 Deep Learning Supercomputer in a Box
#3Any idea how much this costs?
The Nvidia slides had it at $129,000 a pop
Re: The Nvidia DGX-1 Deep Learning Supercomputer in a Box
#4Wow, I didn't realize they were shipping HBM2 already. 720GB/s - with only 16GB of RAM, you can read it all in 22 milliseconds!
Re: The Nvidia DGX-1 Deep Learning Supercomputer in a Box
#5Wait, 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.
Re: The Nvidia DGX-1 Deep Learning Supercomputer in a Box
#6Wait, 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.
Re: The Nvidia DGX-1 Deep Learning Supercomputer in a Box
#7Wait, 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.
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
#8Re: The Nvidia DGX-1 Deep Learning Supercomputer in a Box
#9Check 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 CPU -- of course its faster than a CPU
Re: The Nvidia DGX-1 Deep Learning Supercomputer in a Box
#10Earlier 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
I wish that made that information more accessible. I wasn't able to find it on the site and it was all I really cared about.