News for hackers. Stuff that matters.
Re:The Nvidia DGX-1 Deep Learning Supercomputer in a Box by dazzeruk on Tuesday April 05, 2016 @04:04PM (#11433876)
The Nvidia slides had it at $129,000 a pop
Re:The Nvidia DGX-1 Deep Learning Supercomputer in a Box by rckclmbr on Tuesday April 05, 2016 @04:32PM (#11434113)
Wow, cheap, great for startups!
Re:The Nvidia DGX-1 Deep Learning Supercomputer in a Box by showerst on Tuesday April 05, 2016 @04:39PM (#11434188)
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 by maaku on Tuesday April 05, 2016 @06:40PM (#11435162)
Uh, using what hardware? The 980 Ti is about 11 TFLOP in half-precision (apples to apples). So 16x 980 Ti cards would take up twice as much rack space for $11k. Your estimate (and NVIDIA's pricing) is off by more than an order of magnitude...
Re:The Nvidia DGX-1 Deep Learning Supercomputer in a Box by mon_insider on Tuesday April 05, 2016 @08:39PM (#11435770)
A 980 Ti doesn't have FP16 hardware. The only Maxwell based component with such support is their Tegra part.
Re:The Nvidia DGX-1 Deep Learning Supercomputer in a Box by wmf on Tuesday April 05, 2016 @07:36PM (#11435461)
Isn't the ECC tax around 10x?
Re:The Nvidia DGX-1 Deep Learning Supercomputer in a Box by marshray on Tuesday April 05, 2016 @08:57PM (#11435849)
So, just do the computation twice and compare the results?

OK, so there's twice the power to pay for but it seems like at $129k acquisition cost per 3.2KW consumption you could run for tens of years before break-even.

Re:The Nvidia DGX-1 Deep Learning Supercomputer in a Box by dharma1 on Tuesday April 05, 2016 @07:09PM (#11435320)
I'll take 5