GPU card stuck for all SDK programs


I really don’t know what happened, but suddenly all my cuda program stuck even including those basic ones in SDK, like deviceQuery or clock.
When I run deviceQuery, it just hanging there and showing

[deviceQuery] starting…
./deviceQuery Starting…

CUDA Device Query (Runtime API) version (CUDART static linking)

without any progress to show the real device information. What I did was just changing the cache config and running profiler for a while yesterday. Anyone knows the possible reason for that? or how can I get rid of this trouble? like resetting the GPU card or kill all the task running on that?