Question Deep Learning Workstation Build

Oct 5, 2019
3
0
10
Hi,

My lab group is interested to build a server workstation for our office, a pc which will be used for scientific computing. Here there is a list of components, I think it might suit our needs

Case THERMALTAKE CORE X71 TEMPERED GLASS EDITION GAMING CASE
Custom Liquid Cooling Kit Liquid Series High Kit - EK
Tubing Clear Flexible Tubing (Black Fittings)
Graphics Card Cooling GPU Water Block - For Two Identical Graphics Cards!
Coolant Colour Mayhems Pastel Ice White
LED Lighting 50cm White LED Strip
Overclocked CPU Overclocked Intel® Core™ i9-9940X 14 Core (3.30GHz @ up to 4.4GHz)
Motherboard ASUS® PRIME X299-A: ATX, USB 3.1, SATA 6 GB/s, RGB Ready
Memory (RAM) 64GB Corsair VENGEANCE RGB PRO DDR4 3200MHz (4 x 16GB)
Graphics Card 11GB ASUS ROG STRIX GEFORCE RTX 2080 Ti - HDMI, DP
2nd Graphics Card 11GB ASUS ROG STRIX GEFORCE RTX 2080 Ti - HDMI, DP
1st Storage Drive 1TB SAMSUNG 970 EVO PLUS M.2, PCIe NVMe (up to 3500MB/R, 3300MB/W)

Criticisms, improvements, thoughts welcome!

Thanks,
Andy
 
You also should look at connectivity. Do you have 10GE infrastructure? Are you going to have to move 500GB datasets for training? Is 1TB of total space enough for your datasets ?

And how is the training being stored? Do you need database storage?
 
Hi Kanewolf,

Thank you for help. We have 20TB of shared storage.
But please, let me clarify my question: even if we want to buy a workstation that potential can do machine learning, at the moment we are looking for a system that can run multiple jobs in parallel using CPU and GPU (mostly). I am not an expert and I was not sure about this quote, and so.. I posted here the system build recommended by our supplier. I am sorry I was not clear enough! Thanks!