That's cool. Nearly cool enough to BOIL WATER! After seeing that I think I might upgrade with NVIDIA too. With this >>>
NVIDIA DGX-1
For Christmas this year I would like a deep learning AI computer for my laboratory and the NVIDIA DGX-1 will do just nicely I think.
CPU: Dual 20-core Intel Xeon E5-2698 v4 2.2 GHz
GPU: 8x Tesla GP100 with 128GB of on-board memory
System memory: 512 GB 2133 MHz DDR4 LRDIMM
Hard drive: 4x 1.92 TB SSD RAID 0
Software: Ubuntu Server Linux OS, & NVIDIA Deep Learning SDK, the DIGITS GPU training system, drivers, and CUDA, for designing deep neural networks (DNN)
Power requirements: 3200W (I will need to shut off everything else in the house to keep the electric bill under control)
Hopefully, I can use it to predict protein folding structures and prevent myself and possibly (?) others from ever aging or getting diseases or cancer or diabetes or Alzheimer's/ALS or Parkinson's or Mad Cow and all the different types of Transmissible Spongiform Encephalopathies with a simple pill! Other uses might be generating speech from recordings of people to created synthesized speech making them say what ever I want. Translating an unknown language into any known language. Getting my robot to recognize any kind of object in view, being able to label it all and describe what is happening. Generate unique handwriting styles, also for my robot. Generating poetry and prose as good as any Shakespeare, also for my robot. And having it automatically play an online multiplayer game of DOOM in my place while I am at the grocery store or somewhere else.
CPU: AMD FX-8350 8 core processor 4GHz / GPU: GeForce GT 730 @ 1920x1080, 60Hz with 1GB adapter RAM / RAM: Patriot Signature 4GB 1600MHz 240-Pin DDR3 (only 2GB work, don't buy it) / Motherboard: MSI 970 Gaming MS-7693