Hi,
major issue…!!! I know you can malloc and free memory in kernel. What if you are growing an array, which you do not know its final length?
In these cases i use realloc. However that is not available in CUDA…
Any clues???
Cheers,
Than
Hi,
major issue…!!! I know you can malloc and free memory in kernel. What if you are growing an array, which you do not know its final length?
In these cases i use realloc. However that is not available in CUDA…
Any clues???
Cheers,
Than
Use large enough predefined array?
Regrads,
MK