From: meyerr@cs.ucla.edu (Richard A. Meyer) Newsgroups: comp.parallel Subject: Re: Origin2000 Shared Memory Date: 9 May 1999 04:47:56 GMT Organization: University of California, Los Angeles Approved: bigrigg@cs.cmu.edu Message-Id: <7h341s$j42$1@goldenapple.srv.cs.cmu.edu> Originator: bigrigg@ux6.sp.cs.cmu.edu Xref: ukc comp.parallel:15581 John Roberts writes: >When I run a parallel shared memory (as opposed to MPI) program on an >Origin 2000 how do I specify and determine how many processors I'm >using? What programming environment are you using? A threaded program on a shared memory machine has most of the same properties as an MPI program. You have to create the threads and tell them each what to do. In a C program using Posix threads, for example, you would specify *within the program* how many threads you want to use, by creating that many threads. You could then run 'top' to check your CPU utilization. >What is the difference between each of the following? They all seem to >work to some degree. >setenv NUM_THREADS $(THREADS) >setenv OMP_NUM_THREADS $(THREADS) >setenv MPC_NUM_THREADS $(THREADS) I don't know what any of these are. No doubt they're specific to the programming environment you're using. >If I run 8 threads, what's to say that all 8 threads aren't time >sharing on the same CPU? If you're using the Irix pthread library, you can call to pthread_seconcurrency(N) to try to force the system to use N processors. By default, threads may share CPUs. >If I get 8 CPU's do I have exclusive access to them, No. > or can other threads time-share with mine? Yes. > Can my threads be migrated to other processors? perhaps, but we have not experienced this that I know of. Windows NT does this routinely. It's very annoying. > How do I know when this happens? You probably wouldn't. As a shared memory machine, the threads should be transparently portable between processors with only a temporary slowdown in performance. Rich Parallel Computing Laboratory, UCLA -- Articles to bigrigg+parallel@cs.cmu.edu (Admin: bigrigg@cs.cmu.edu) Archive: http://www.hensa.ac.uk/parallel/internet/usenet/comp.parallel