Deep Learning Build - GPU Manufacturer Matter?

[PikaNim]

Hi, I’m relatively new to the PC building world. And I’m looking to get into deep learning. Based on my budget, the GTX 2070 Super is the best I can afford. However, i wasn’t sure if the manufacturer of the GPU matters for deep learning? Do  have to get the base model by NVIDIA? Or are the other manufacturers like ASUS, MSI, EVGA, etc an option? Could someone tell me if there is a difference? 

[Fasauceome]

The difference is in the cooler design, and for the high end, power delivery. Basically, you'll have the same amount of tensor cores no matter what 2070 super you get.

[boggy77]

i would opt for a design with three fans, they usually cool better, so they are quieter.

[Wh0_Am_1]

4 minutes ago, PikaNim said:

Hi, I’m relatively new to the PC building world. And I’m looking to get into deep learning. Based on my budget, the GTX 2070 Super is the best I can afford. However, i wasn’t sure if the manufacturer of the GPU matters for deep learning? Do  have to get the base model by NVIDIA? Or are the other manufacturers like ASUS, MSI, EVGA, etc an option? Could someone tell me if there is a difference? 

Welcome to the LTT forums! To answer your question generally speaking no, as they all include the same basic Nvidia components Nvidia's board partners just stick another cooler or modify the board at times to improve the performance of the card. Personally I would recommend going EVGA as they tend to have the best warranties.