Thursday, November 17th 2016

NVIDIA Announces DGX SaturnV: The World's Most Efficient Supercomputer

This week NVIDIA announced their latest innovation to the HPC landscape, the DGX SaturnV. Destined for the likes of universities and companies with a need for deep learning capabilities, the DGX SaturnV sets a new benchmark for energy efficiency in High Performance Computing. While not managing the title of the fastest supercomputer this year, the SaturnV takes a respectable placing of 28th in the top 500 list, while promising much lower running costs for performance on tap.

Capable of delivering 9.46 GFLOPS of computational speed per Watt of energy consumed, it bests last years best effort of 6.67 GFLOPS/W by 42%. The SaturnV is comprised of 125 DGX-1 deep learning systems, and each DGX-1 contains no less than eight Tesla P100 cards. Where a single GTX1080 can churn out 138 GFLOPS of FP16 calculations, a single Telsa P100 can deliver a massive 21.2 TFLOPS. The singular DGX-1 units are already in the field, including being used by NVIDIA themselves.
Add your own comment

5 Comments on NVIDIA Announces DGX SaturnV: The World's Most Efficient Supercomputer

jabbadap said:
DGX SaturnV is nvidia's in house super computer, I don't think they will start make and sell those outside(Or will they?).

Cray's xc50 uses tesla P100 too, and is in use on Piz Daint super computer(#8 on top500).
like all such machines they are built as ordered. the market for them is very small. I wouldn't expect to see it on CDW.
Posted on Reply
Time to buy 126 units and take their spot on the list!

Not that I have any specific plans for that much computational power... or know how to pay for it all. Would be $16m+ for 126 DGX-1s, then ya gotta rack them (12), power them (400k+ watts), network, etc. Hmmm..... maybe I'll pass.
Posted on Reply
So it's a 21.2 PFLOPS machine according to your calculations (125*8*21.2 TFLOPS) :huh:
Posted on Reply