Run on dione

[RayChromium]

First of all, now it can be compiled and run on dione after fixing some errors;
and I tried to eliminate the loop in calculateHistograms by mapping the one-dimensional thread ID onto two-dimensional indices. A common way (from good old google 🙏) to do this is to use division and modulus operations, here I did i = index / numD and j = index % numD map the one-dimensional thread ID onto a two-dimensional pair (i, j).
Let's say the maxInputLength is 4, then all the mapped indices are:

(0, 0), (0, 1), (0, 2), (0, 3),
(1, 0), (1, 1), (1, 2), (1, 3),
(2, 0), (2, 1), (2, 2), (2, 3),
(3, 0), (3, 1), (3, 2), (3, 3)

And it seems to work.

Also, modify the printf with outfil pointer using fprintf, so that the result is saved in omega.out.....but the results looks a little bit wierd though......

haven't got idea about the wierd values in omega.out at lines around 20.000, the omega values are so large around 1000, don't know it that make sense.

[RayChromium]

But at least it gets the job done, good "start" though...if this make sense, now we can think about how to improve the runtime I guess

[RayChromium]

And also, I figured out how to build and run this on dione after its upgrade:

1. when login to dione, run module load gcccuda and module load cuda, these are the modules that actually works;
2. go to the directory of this repo, run ./generate.sh make and ./generate.sh run, I simply wrapped the building commands in the makefile and the run command in the shell script