[mpich-discuss] MPI_Waitany: "The semaphore timeout period hasexpired. ( errno 121)"

Jayesh Krishna jayesh at mcs.anl.gov
Tue Jun 2 09:48:20 CDT 2009


Hi,
 
# Are you using MPICH2 on unix or windows ?
# What is the complete error message that you get (copy-paste the complete
error message)?
# Do you have a sample test program showing the problem ?
 
(PS: As Anthony mentioned in his email look into MPI dynamic process
management for dynamically managing MPI processes -
http://www.mcs.anl.gov/research/projects/mpi/mpi-standard/mpi-report-2.0/n
ode90.htm#Node90)
 
Regards,
Jayesh

  _____  

From: mpich-discuss-bounces at mcs.anl.gov
[mailto:mpich-discuss-bounces at mcs.anl.gov] On Behalf Of
David_Lowinger at ea.epson.com
Sent: Monday, June 01, 2009 5:23 PM
To: mpich-discuss at mcs.anl.gov
Subject: [mpich-discuss] MPI_Waitany: "The semaphore timeout period
hasexpired. ( errno 121)"



I am encountering the following error at runtime when using MPI_Waitany: 

"The semaphore timeout period has expired. ( errno 121)" 

It looks like it's occuring in MPIDU_Sock_wait around line 2603.  Can I
get more information on this error somewhere?  Is there a way to increase
this "timeout period"? 

On a separate note, is there a proper way to implement a "dynamic network"
using MPI, ie. a network in which nodes are added or subtracted during
runtime?  Can someone recommend a good document or method for that? 

Thank you and regards, 
David 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/mpich-discuss/attachments/20090602/9aca934e/attachment.htm>


More information about the mpich-discuss mailing list