@Axel Vogt Thanks for all the usefull advices guys now I am modifying all of the procedures to be suitable to compiling. Hope I'll find this SNAIL in my code...when I do it I'll write for sure ;-).

@Axel Vogt Thanks for all the usefull advices guys now I am modifying all of the procedures to be suitable to compiling. Hope I'll find this SNAIL in my code...when I do it I'll write for sure ;-).

@Axel Vogt Actually there are many procedures I have generated and I don't know where could be the keystone of the speed up, but for sure I know there is some BUG slowing down the process. If you could give me some strategy of searching then just using the profiler and TIME function of maple I'll be glad.

@Axel Vogt Actually there are many procedures I have generated and I don't know where could be the keystone of the speed up, but for sure I know there is some BUG slowing down the process. If you could give me some strategy of searching then just using the profiler and TIME function of maple I'll be glad.

@Axel Vogt Ok so I actually use MAPLE to generate a program for modeling of boundless domains subjected, to dynamic excitations. Using the frequency content of certain excitation I need to generate the dynamic stiffness matrix (Impedance matrix) of a boundless elastic domain. Ofcourse I use boundary elements technology for that and so I generated a code which works satisfactory with respect to results...but unfortunately poorly with respect to time...so briefly now I am desperately trying to speed up the procedures.

@Axel Vogt Ok so I actually use MAPLE to generate a program for modeling of boundless domains subjected, to dynamic excitations. Using the frequency content of certain excitation I need to generate the dynamic stiffness matrix (Impedance matrix) of a boundless elastic domain. Ofcourse I use boundary elements technology for that and so I generated a code which works satisfactory with respect to results...but unfortunately poorly with respect to time...so briefly now I am desperately trying to speed up the procedures.

Wow you are really REALLY GOOD man, thats absolutely correct...bytheway I managed to generate a relatively fast procedure, close to that you gave me and it could be compiled to, I'll show you soon. I actualy used this one because maple Bessel functions are really expensive...can you actually tell me how to compile them if it's possible. The autor is J. Dominguez...it's about boundary elements method in dynamics.

Wow you are really REALLY GOOD man, thats absolutely correct...bytheway I managed to generate a relatively fast procedure, close to that you gave me and it could be compiled to, I'll show you soon. I actualy used this one because maple Bessel functions are really expensive...can you actually tell me how to compile them if it's possible. The autor is J. Dominguez...it's about boundary elements method in dynamics.

If I am wrong about that...could you PLEASE tell me what are the reasons why increasing GCFREQ leads to increased computational speed?

@acer I have the array generated in maple enviroment and the size of the array is just say RESPECTABLE and I think that garbage collection scans the array for RESPECTABLE amount of time too. I think gc slows down my computations because when I increased kernelopts[gcfreq]=5000000 (not the default 1000000) the computational speed increased noticeably BUT NOT ENOUGHT. So if I am not wrong gc is doing something with the large amount of data that I wont alter anyway...so I don't want this data to be BOTHERED...hope I am clear. And yes OpenMaple procedures can't help because I am not using external code.

Wow Axel that answer of your really rocks...thanks man. You've speeded my code with years.

Anyway my type of XIn is complex[8], modulus is for sure greater then 0 and not equal to 5, as for the global arrays of numbers...I took them from a code in a book. But why you are asking abut the decimal places, how much should they be?

I'll be glad if you got something more, but I am afraid that the time can take the negative sign then :-). Thanks for the advices guys, this realy improved the performance of the code. When I got 10 pts of reputation I'll tumb up for sure.

Wow Axel that answer of your really rocks...thanks man. You've speeded my code with years.

Anyway my type of XIn is complex[8], modulus is for sure greater then 0 and not equal to 5, as for the global arrays of numbers...I took them from a code in a book. But why you are asking abut the decimal places, how much should they be?

I'll be glad if you got something more, but I am afraid that the time can take the negative sign then :-). Thanks for the advices guys, this realy improved the performance of the code. When I got 10 pts of reputation I'll tumb up for sure.