> > I am trying to find out how many copies of ScratchData and CopyData will
> > be generated in WorkStream::run(). The documentation seems to be silent
> > about this.
> 
> It seems to be
>   queue_length * chunk_size * chunk_size
> given the arguments to WorkStream::run(). I must admit that I now think the 
> second chunk_size factor is probably a mistake. I'll check this out.

Taking into account the change Wolfgang made yesterday, a more precise
answer for the number of copies in the current implementation: 
- ScratchData: queue_length
- CopyData: queue_length * chunk_size

This is because the scratch data is supposed to be reused while working
on a chunk of iterators (cells) and filling up the copy data.

Best,
Martin

_______________________________________________
dealii mailing list http://poisson.dealii.org/mailman/listinfo/dealii

Reply via email to