Calculation taking too long, stats from scipy may be the problem?
Hey!
I have a question regarding a calculation I'm running. The code includes only 2 loops, one inside another, and some calculus inside. The loop run over many iterations (~100000X100000), but it is not an excuse for the calculation to be taking (so far...!) over 20 hours!
- The calcultation inside the loops is a fairly simple one: W[k]=W[k]+1/2...+..., but - i also include - sqrt(bla bla bla)stats.norm.rvs . One of the reasons for the ridiculously long calculation time, as I heard, might the calling for library from Scipy, where - as it means calling an outside library, which I don't know how it works, increeses the time of every step in the calculation in such a way that for such many iterations results in the long running time.
Can this be the source of the problem? Can calling an outside library, no matter if - Scipy of whatever else, increase the calculation time by so much?
What can I do?
Thanking you on advance!
- Also, can it be that because Sage and Python are not as basic as programming languages such as C and Fortran etc. - every calculation in them is always bound to take much longer time then in the others??
Well, it depends - a lot of stuff interfaces directly with C libraries, and Cython stuff is really more or less as fast as C.
More to the point, we really need to see the *specific code* in order to profile this. There are also profilers such as %prun on the command line - and see http://ask.sagemath.org/question/693/profiling-code-in-the-sage-notebook as well. But without the code it's impossible to say.
O.K, thanks man
But you should feel free to edit your original question to show that code! Someone may then be able to help.