Re: [bgl-discuss] error messages

[Stephen Siegel <siegel@xxxxxxxxxxxx>]

I believe I am just using the defaults at all stages (compiling,
linking, and executing).  Specifically:

COMPILER   = blrts_xlc -I/bgl/BlueLight/ppcfloor/bglsys/include
LINKER     = blrts_xlc -L/bgl/BlueLight/ppcfloor/bglsys/lib
LIBRARIES  = -lmpich.rts -lmsglayer.rts -lrts.rts -ldevices.rts

and to execute:

cqsub -q short -t 30 -n 32 ...

I should add that as an experiment, I tried re-running the program
after changing just a single parameter which has the effect of
reducing the size of all memory blocks by a factor of 10, and hence
reduced the size of all messages by a factor of 10, but left the
entire communication pattern unchanged.  The program ran successfully.
So it does seem to be a memory problem.

-Steve




On Sat, 21 Jan 2006, Steven Pieper wrote:

> >>> I was wondering if anyone could help me interpret the error messages
> >>> Rzv:cannot allocate unexpected buffer from R:30 T:0 C:124
> >>> 
> 
> Sometime ago I was getting the "unexpected buffer" messages in a Fortran
> program that was sending lots of messages to the master node from many
> slaves.  I switched to usisng mpi_ssend which fixed the problem (and
> seemed to actually run faster in some cases thta had not failed).  The
> messages I got were not at all as informative as yours, basically just
> cannot allocate.  were you running with some higher than default debug
> level?
> 
> Steve
> Steven C. Pieper, Senior Physicist:  spieper@anl.gov  
> Argonne National Laboratory, Physics Division, Bldg. 203, Argonne, IL 60439
> Phone:  630-252-4232         Fax -6008
> Secretary (Donna Nelson or Debra Morrison) -4100
> 

- --------------------------------------------------------------------
To add or remove yourself from this mailing list, use the 'notifyme'
command on any BGL machine. To remove: notifyme -n, to add: notifyme -y.