Hi there ,
I am trying to use my Tesla c 1060 GPUs for deep learning application in Matlab 2017b and I had faced a problem which is this version of Matlab requires a graphics driver that support CUDA 8 or above. the most recent driver for Tesla C 1060 is supporting CUDA 6.5, so i have got an advice from Matlab community to find a compatible 37x driver for my tesla c1060 that supports CUDA 8.
Is there any other GPU driver that might be Ok to use so I can later on install CUDA 8, for instance I have tried the driver of c2050 but unfortunately it didnt work.
All the deep-learning frameworks I have come across require compute capability >= 3.0. The Tesla C1060 has compute capability 1.3, so what you are trying to do will not work. No version of CUDA past CUDA 6.5 will run with this GPU; the currently shipping version of CUDA is 9.0.
A reasonably priced modern consumer GPU like the GTX 1050Ti will provide the same amount of memory (4 GB) and better performance than the C1060 at 35% of its power consumption and lets you use the latest version of CUDA and the latest deep learning frameworks.
IMHO, the best place for a C1060 at this time is an approved e-waste recycler.