Comment Re:Tied to a card (Score 1) 134
That's where abstraction and specialization comes into play. After defining your algorithm for independent use, specialize and optimize it to exploit current or future hardware. This gives you a fallback for calculation, and extremely enhanced performance for the life and support of said hardware. And, as others have pointed out, it's a stepping stone to an OpenCL implementation, eventually giving you multiple vendors to rely on.
If NVIDIA goes out of business or drops support in two years, how much more work will you have gotten done over that time? If it's any less than the cost of implementing the specialized solution, it's worth it.
Is there risk? Yes. And, it's highly mitigated with the abstracted solution and migration paths.