requirements for upgrading 1080ti for deep learning

May 1, 2018
1
0
10
hi, i am thinking to upgrade my GPU to 1080ti. I am gonna use it mainly for deep learning, not much in gaming (nearly none). is there a better choice of GPU with similar performance but lower price or that's the best one? If I decided to get 1080ti, what is the minimum changes for my other components of my build to make it run, like do I need more Watts? do 1080ti fits the slot in the motherboard? Also, would there be a bottleneck of performance from any of my other components? Thank you

Here's my current build:
CPU: i5 4440 @ 3.1GHz
GPU: Nvidia GTX 660
Motherboard: Gigabyte GA-B85M-D2V
RAM: HyperX Fury DDR3 1600 C10 2x8GB
Memory: SSD 120GB, HDD 1TB
Power Supply:Thermal Master ATX 12V (420W)
 
Solution


I'd look a little deeper into Vega 64...


I'd look a little deeper into Vega 64 and its benefits. If you do purchase a 1080 Ti I recommend the EVGA 1080 Ti FTW3 Hybrid. A quick Google(I know not of that which you speak) reveals https://wccftech.com/amd-radeon-vega-nvidia-pascal-gp100-performance-benchmarks/.

The 4790 might not be too expensive in the used department.

I chose the Seasonic G-Series 650W 80+ Gold PSU. It was $20 cheaper than their Focus model which is a popular choice these days at that wattage point. The Corsair Rx650m is another good choice but again it's expensive. All 1080 Ti's call for a 600W or better PSU iiinm.
 
Solution