[Scilab-users] Duration time of Xcos simulation

[Serge Steer]

It should be better to replace the “sci_func_block_m”  by the 
“EXPRESSION” one.

Moreover in your case I think that the main part of the simulation time 
is due to diagram compilation,  if you need to perform several 
simulations of the same diagram (up to context changes) you must 
re-enter the info parameter of scicos_simulate to vaoid recompilation 
for each simulation.

Note also that the the sampling time you enter is in fact the maximum 
step size the solver can use. It may take smaller steps to achieve the 
requested precision.

Serge Steer
INRIA
Le 18/07/2013 12:52, simi99 a écrit :
> Good afternoon everybody,
> I simulate a servo position (cart) in Scilab / Xcos. The model equation
> (differential equation) is implemented by “sci_func_block_m” in Xcos.
> RK45-Runge-Kutta solver is set with a sampling time of T = 1ms. The
> simulation time is Tsim = 5sec. The simulation data is carried to Scilab by
> a "ToWorkspace"block. (There are no Scope blocks in the schematic) The
> simulation takes an average time of 1.8-2 seconds, which is not acceptable
> because I run the simulation in batch mode. To comparison: Matlab/Simulink
> does the same job (same boundary conditions) in about 100ms. Is there
> something I can do to improve the performance? Thanks in advance,
> Simon
>
> <http://mailinglists.scilab.org/file/n4027041/PosServoScilabMailingList%281%29.png>
>
> <http://mailinglists.scilab.org/file/n4027041/PosServoScilabMailingList_Solver.png>
>
>
>
> --
> View this message in context: http://mailinglists.scilab.org/Duration-time-of-Xcos-simulation-tp4027041.html
> Sent from the Scilab users - Mailing Lists Archives mailing list archive at Nabble.com.
> _______________________________________________
> users mailing list
> users at lists.scilab.org
> http://lists.scilab.org/mailman/listinfo/users