When using the
nvprof tool to profile Numba jitted code for the CUDA
target, the output contains
No kernels were profiled but there are clearly
running kernels present, what is going on?
This is quite likely due to the profiling data not being flushed on program
exit, see the NVIDIA CUDA documentation for
details. To fix this simply add a call to
to the exit point in your program (or whereever you want to stop profiling).
For more on CUDA profiling support in Numba, see Profiling.