I am planning a build for machine learning purposes, more specifically, training neural network models for image classification/localization etc, budget is below 2000 euro.
Provisional build is listed here:
PCPartPicker part list: https://pcpartpicker.com/list/MW8PHN
Price breakdown by merchant: https://pcpartpicker.com/list/MW8PHN/by_merchant/
CPU: Intel Xeon E3-1231 V3 3.4GHz Quad-Core Processor ($239.99 @ SuperBiiz)
CPU Cooler: Corsair H100i 77.0 CFM Liquid CPU Cooler ($99.99 @ Corsair)
Motherboard: MSI Z97-GAMING 5 ATX LGA1150 Motherboard
Memory: Crucial Ballistix 8GB (1 x 8GB) DDR3-1866 Memory ($63.98 @ Directron)
Memory: Crucial Ballistix 8GB (1 x 8GB) DDR3-1866 Memory ($63.98 @ Directron)
Storage: Intel 600p Series 512GB M.2-2280 Solid State Drive ($169.99 @ SuperBiiz)
Video Card: EVGA GeForce GTX 1080 8GB Superclocked Gaming ACX 3.0 Video Card ($499.99 @ B&H)
Case: NZXT S340 (Black/Red) ATX Mid Tower Case ($64.99 @ Newegg)
Power Supply: EVGA SuperNOVA NEX 650W 80+ Gold Certified Fully-Modular ATX Power Supply ($74.89 @ OutletPC)
Total: $1277.80
I have a few questions in mind:
I have opted for a Xeon CPU without integrated graphics, because I already have a GPU on board. Will that make it difficult for me to install an OS (Ubuntu Linux) on the machine? I am a bit concerned about the "you need to install a driver to display things, but you need to display things in order to install the system to begin with" catch-22.
I don't need a gaming/fancy motherboard since this is supposed to be just a number cruncher. In fact, I don't even need a keyboard or monitor once the system is setup: I can just ssh into it and execute programs. So can I save a bit money here without compromising performance?
Does the power supply listed above have enough wattage to support the whole build? What if I want to add more GPUs to it in the future?
Any other suggestions will also be welcome.