mpi conditional point to point communication
in [Fortran]
|
Prev: How to pass function names as arguments when function returns an array?
Next: Interpolation between Bezier curves
From: bingo on 17 Feb 2010 14:30 Thanks a lot for the reply. But in my understanding, all the collective communications like mpi_gather require synchronization, correct? My problem really do not have this feature at all. Maybe I'm just being stupid, but it would be nice if the rank=1 processor can communicate with master node with no dependence on the status of other slave processors. On Feb 16, 1:23 pm, Harald Anlauf <anlauf.2...(a)arcor.de> wrote: > On Feb 16, 9:50 pm, bingo <zhn...(a)gmail.com> wrote: > > > > > Now I have another question. By doing the loop, as shown below, I'm > > actually applying dependency between individual slave processors. That > > said, now the rank=2 need to wait for master processor to receive > > result from rank=1. Even though the calculation on rank=2 might finish > > earlier, it won't receive further data until the calculation on rank=1 > > is finished. Even worse, the whole process won't proceed any further > > if the calculation on one processor get stuck. I think this should be > > a problem of my bad coding algorithm. Ideally, the slave processors > > should ran independently, and only communicate with the master node. > > Is there a way to improve on this? > > > Thanks again. > > Bin > > > ===================================== > > do rank = 1, nprocs-1 > > ! send initial condition > > call mpi_send(config, 4, mpidp, rank, confTag, > > mpi_comm_world, ierror) > > ! receive result > > call mpi_recv(endInfo, 7, mpidp, rank, endTag, > > mpi_comm_world, istatus, ierror) > > enddo > > Well, this is not really a Fortran-related question, > but anyway, you have multiple options: > > First split the loop into parts where initial data are > distributed, then another loop where results are collected. > In these loops: > > - Either use mpi_scatter/mpi_gather to distribute the data to > the individual PEs, or receive the results. (Recommended) > > - use explicit coding instead of the collective MPI operations. > > E.g.: > > do rank = 0, nprocs-1 > if (node == 0) then > if (rank /= node) then > call mpi_send (..., rank, ...) > end if > else ! node /= 0 > if (rank == node) then > call mpi_recv (..., 0, ...) > end if > end if > end do > > to distribute the data to the PEs, and corresponding code > to gather the results. > > (Warning: the above code has not been tested, but I think you > get the idea. I use similar code on machines where mpi_scatter > has poor performance or when I need to debug.)
From: Victor Eijkhout on 17 Feb 2010 15:54 bingo <zhngbn(a)gmail.com> wrote: > Even though the calculation on rank=2 might finish > earlier, it won't receive further data until the calculation on rank=1 > is finished. Yep, I noticed that. You need to look into the Isend and Irecv calls. They send the data off asynchronously. Then when you've sent data to all slaves, do a receive from MPI_ANY_SOURCE and process data as it comes in. Victor. -- Victor Eijkhout -- eijkhout at tacc utexas edu
From: Harald Anlauf on 17 Feb 2010 15:56
On Feb 17, 8:30 pm, bingo <zhn...(a)gmail.com> wrote: > Thanks a lot for the reply. But in my understanding, all the > collective communications like mpi_gather require synchronization, > correct? My problem really do not have this feature at all. Maybe I'm > just being stupid, but it would be nice if the rank=1 processor can > communicate with master node with no dependence on the status of other > slave processors. I spend quite some effort on getting a good load balance of my application so that I can use blocking communication and get good efficiency. From your last comment I infer that your application should dynamically distribute the load. I think you're better off asking in an MPI- related newsgroup. |