I have been using Nvidia cards since the original geforce 256 in 2000. I purchased one of the first ATI 8514 Ultra cards in the early 90s. I own two R9-290s, and at work I use Tesla K40s. I follow GPU computing avidly as I do machine learning using mainly openCL (for my sins), precisely because I want choice in the market. It would be much easier for me to do CUDA. I resist for reasons of rejecting proprietary APIs.
I know exactly what I am talking about. I don't forgive Nvidia its sins, I actually prefer AMD, but I am not about to lie to myself that Nvidia is somehow an unsuccessful company, when it is worth 7x more than AMD, and AMD also includes an X86 line. It's a no-contest scenario in the eyes of the market, much to my chagrin, but I cannot deny the reality:
