I’m developing with NVIDIA’s XAVIER. I plan to implement cuFFT using CUDA, get a profile and check the performance with NVIDIA Visual Profiler.
I will paste part of the source code and the result of profiling it with nvprof.
Input data of 256x64 is read from Excel (omitted), and it is calculated by cuFFT.
so I have questions about nvprof result.
void regular_fft<unsigned int=64, unsigned int=8, unsigned int=32, padding_t=1, twiddle_t=0, loadstore_modifier_t=2, layout_t=1, unsigned int, float>(kernel_arguments_t)
void vector_fft<unsigned int=256, unsigned int=16, unsigned int=1, padding_t=6, twiddle_t=0, loadstore_modifier_t=2, layout_t=0, unsigned int, float>(kernel_arguments_t)
What does the above mean?
For example, padding_t = 6, twiddle_t = 0
About the meaning of the value of.