why CUDA7.5 instead of 8.0?
I got a big request from nvOC users, Ubuntu 16.04-based, all I could get out-of-the-box for this OC was CUDA 7.5. I'm quite busy (and seems that I'll be busy for a long time) with current tasks from my primary work, I just filled a big community request and got back to my primary tasks.
Maybe I'll update it once more when receive feedback from my tester, another version is ready to deploy, but needs testing.
Thank you for understanding.