Beostatus and Scyld Configuration Question

Michael Wiltberger wiltbemj at
Mon Apr 23 15:59:54 PDT 2001

  I've just used the Scyld Beowulf 2.0 Preview release to setup a 13
node 25 processor beowulf cluster.  Each of the 12 compute 12 nodes has
128MB of memory as well as a 4GB disk drive and has been setup using the
default options.  When I run beostatus I see that approxmately 95% of
the memory is currently in use.  Is bproc really using 121MB of memory? 
When I try to run a simple code which allocates 120MB of memory I get
the following error from mpi
p0_23484:  p4_error: interrupt SIGSEGV: 11

Have I misconfigured something or do we need more memory for the
system.   Any advise is appericated.

Take Care,

|Michael Wiltberger, Ph.D     |                       |
|Research Asst. Professor     | Life's unfair -       |
|Dartmouth College            |  but the root         |
|Dept. of Physics & Astronomy |  password helps!      |
|6127 Wilder Laboratory       |                       |
|Hanover, NH 03755            | VI VI VI              |
|(603)646-0428                |   The editor of the   |
|wiltbemj at|   beast               |
|www.myst.dartmouth/wiltbemj/ |                       | 

More information about the Beowulf mailing list