Sobol First order and total sensitivities

I have 19 input variables and two outputs. I am focusing on one at a time. I have a closed form model flike Y = F(x1, x2, x3.... x19) for example so the function values can be computed almost instantly oncethe input vector X is known. I am using some program GSAT frome File exchange. However, I realizedthe number of combinations it has to compute is way too high like 500,000+ or so. My program has been running already more than a day and I am not sure what is going on. The same program worked fine when I had 9 inputs (511 combinations) and a simple model and result came out like less than an hour. My question is, whether I am doing something fundamentally wrong and there might be a better way or not? please help. Thanks in advance.

[John D'Errico]

I took a look at GSAT. My initial guess would always be the code was just written poorly and inefficiently, but the code seems to have been written by someone who probably knew what they were doing. Could it have been done more efficiently? Certainly so, since any substantial piece of code has room for improvement. But I'd not quibble there.

Anyway, option 2 is just more likely. People solve small problems, and find they work quickly. So the answer is to throw a much larger problem at the code, and not realize the complexity grows rapidly with problem size.

I would just suggest you accept that big problems require big, FAST computers, or a large cup of coffee and a willingness to sit down with a large book to read while you wait for a result. The complete works of Shakespeare might be a good start. Don't forget to read all the footnotes. And read it a second time, because you probably missed something the first time through.

[Pappu Murthy]

I understand what you are saying. My problem is perhaps too big to handle for a PC at home. But I needed an independant confirmation which you gave. Thanks. But let us see for more answers and if anyone else have any ideas.