← Back to team overview

yade-dev team mailing list archive

Re: 3 x speedups with new containers and openMP!

 

> For BEX & VecMap & openMP options (but without new InteractionDispatchers yet):
> for STLImporterTest.py with 60438 spheres, 
> 1 thread:  15280 iters / 16632 sec = 0.91 iter/s, 50% CPU usage
> 2 threads: 21173 iters / 18820 sec = 1.12 iter/s, 65-70% CPU usage
> So, speedup is more 20% with 2 threads for big simulations. Good!

with InteractionDispatchers:
2 threads: 21173 iters / 16668 sec = 1.27 iter/s,
speedup is 13%.

-- 
Best regards, Sergei D.



References