What would be a good GPU to start learning CUDA on?
Are any GPUs unsuitable?
closed as not constructive by Steve Evers, gnat, Walter, Yusubov, StuperUser Oct 17 '12 at 12:38
As it currently stands, this question is not a good fit for our Q&A format. We expect answers to be supported by facts, references, or expertise, but this question will likely solicit debate, arguments, polling, or extended discussion. If you feel that this question can be improved and possibly reopened, visit the help center for guidance. If this question can be reworded to fit the rules in the help center, please edit the question.
Virtually any card made by nVidia in the last 5 years can be used to run CUDA programs. In terms of "Best", you'd probably want a Tesla but they are hideously expensive. By hideously expensive, I mean that they are designed to operate only as computing on GPU hardware and not as regular graphics cards. You mount them in servers and "Do Science" to things with them. You can technically mount one in a workstation but it's still expensive. A bare minimum machine from Dell, for example, is just under $5000.
As far as home applications, whatever card you've got would probably work fine provided that it's on this list. Pick something you can afford. Find some problems that are of interest and amenable to the SIMD(SPMD) architecture and have at it. Finding a reference for C is recommended as well as a CUDA guide. If you don't yet know any C, Zed Shaw will cure you of that.