> > Have anyone ever run a simulation with 100,000 spheres? I run one with my > desktop and it takes forever for one loop. I let it run 4 hours and no thing > happen (in null GUI). My desktop is AMD 6400, 2gb RAM, > Hi, I was recently running (for speed measurements) TriaxialTest with 40000 elements. It took maybe 5 minutes before the simulation started, then it ran at steady speed. The figure linked below shows #elements and time per iteration graph, it is nicely linear, so I would expect about 1 iteration / sec with 100000 spheres at my computer (slower than yours). RAM requirements are negligible, I didn't notice any bump on the RAM monitor in the panel. beta.arcig.cz/~eudoxos/temp/[EMAIL PROTECTED]
I didn't investigate what was the cause of the lag at the beginning - either root body initializers or first run of the SAP collider that has to fill its bound containers. I may profile that with kcallgrind, but I fear in that case those 5 minutes would take forever. Regards, Vaclav _______________________________________________ Yade-users mailing list [email protected] https://lists.berlios.de/mailman/listinfo/yade-users
