On 04/03/2014 05:36 AM, Biddiscombe, John A. wrote:
Greetings
I’ve been forwarded some info regarding a hang on hdf5 read from a user
of enzo.
The problem appears when 512 (or more) processes open a file in serial
and each independently read a hyperslab (non overlapping, I believe, but
this may not be correct).
I asked for a stack trace and the attached screenshot was sent to me.
darn, the call graph is truncated just as it gets interesting.
Is there a chance that it is not hung, but instead it is progressing
really really slowly? for certain data descriptions, HDF5 or MPI-IO
might be doing element-at-a-time i/o. Or it might be doing something
like data sieving where it will read a lot of data in order to get to
the few items of interest.
==rob
--
Rob Latham
Mathematics and Computer Science Division
Argonne National Lab, IL USA
_______________________________________________
Hdf-forum is for HDF software users discussion.
[email protected]
http://mail.lists.hdfgroup.org/mailman/listinfo/hdf-forum_lists.hdfgroup.org
