add tuning option to all kernels, provide _tune functions for all kernels in variadic format and completing the lua file
I provide a new version of lua work in this commit. I hope it would be fine.
Also two issues are in my mind:
1) It would be nice if the real (tuned) value of nb is printed in results. Already, the printed value is set equal to 256.
2) In zgbtrf.c, there is a quick return comment without any code under it. It probably needs to be edited.
Thank you so much for your time and concern