Hi, Wolfgang, Thank you for your reply and suggestion. Your're right, I didn't conduct profiling yet. My commonly used test setup is 1 node (128 cores) with 128 tasks on it, 99% of assembly time came from this type of run. It's a bit surprising for me to be suggested not to use thread, but I believe it should be a significant experience. I will do profiling first and then talk in this thread later.
Sincerely, Timo On Fri, Sep 15, 2023 at 3:51 PM Wolfgang Bangerth <bange...@colostate.edu> wrote: > On 9/15/23 03:17, timo Hyvärinen wrote: > > > > So here is my question, which tutorial steps demonstrate how to > implement the > > mpi-thread hybrid parallelism. I've found step-48 is talking about this, > but > > I wonder are there any other tutorial programs to look at? I also wonder > if > > any of you guys have suggestions about mpi+thread parallelism under > > dealii framework? > > Timo: > If your code already works, then the usual suggestion would be to use one > MPI > process per core, and to not use thread-parallelism at all. In that case, > you > should be using all cores equally. > > There is an often-repeated observation that one can not optimize a code > without first profiling it. From your question, it sounds like you are not > entire sure why assembly is taking so long. I think it is likely that > whatever > solution you try will not be successful unless you first find out why it > really is taking this long. I would run the code on one MPI process first > and > see whether assembly is taking 99% there as well. If that's the case, use > valgrind's callgrind to figure out why. If it's not taking 99% on one MPI > process, then you've got another riddle to solve. > > Best > W. > > -- > ------------------------------------------------------------------------ > Wolfgang Bangerth email: bange...@colostate.edu > www: http://www.math.colostate.edu/~bangerth/ > > > -- > The deal.II project is located at http://www.dealii.org/ > For mailing list/forum options, see > https://groups.google.com/d/forum/dealii?hl=en > --- > You received this message because you are subscribed to the Google Groups > "deal.II User Group" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to dealii+unsubscr...@googlegroups.com. > To view this discussion on the web visit > https://groups.google.com/d/msgid/dealii/140169bb-1eb5-dcfd-44ef-14081aa713c5%40colostate.edu > . > -- The deal.II project is located at http://www.dealii.org/ For mailing list/forum options, see https://groups.google.com/d/forum/dealii?hl=en --- You received this message because you are subscribed to the Google Groups "deal.II User Group" group. To unsubscribe from this group and stop receiving emails from it, send an email to dealii+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/dealii/CAArwj0HQnSVDSXmsh6CmdtDW7Y2zNF_jagbo-hJC_btg7FuuFA%40mail.gmail.com.
