Which GPU is better for Deep Learning GTX980M or GTX1060

I’m a newbie to deep learning and I have a plan to start using deep learning in wireless communications, but I need to start with simple models of deep learning to digest the ideas and the principles of deep learning. My first step is to buy an affordable laptop but I’m confused a little bit about which GPU is better: GTX 980M or GTX1060. I found that GTX980M has good memory bandwidth is(160GB/s), the number of cores is 1536 while the VRAM is 8GB while GTX 1060 has higher memory bandwidth (192GB/sec) with less number of cores (1280) and less VRAM 6GB.
What does make a GPU better than another one?

1 Like

These are pretty close options and either will get you started. A little speed vs memory tradeoff, but I did my initial learning on the 1060 and did fine with the limited memory. My daughter has the 980 and has tried a bit of DL on it. Don’t sweat this choice - pick one and get going! You can always spin up a cloud instance when your model is too hefty for your laptop.


I think GTX 1060 is a better choice as its new gen card and will help in the long run