HP-MPI Version 2.2.7 for Linux Release Note

I is 0,...,ncard-1
k is 0,...,nport[next_card]-1
m is card count so far
1.2.7.14 Expanded lightweight instrumentation functionality
This release extends the HP-MPI lightweight instrumentation feature in a number of ways.
Instrumentation is now supported when using -ha. If some ranks terminate during or prior to
MPI_Finalize(), then the lowest rank id in MPI_COMM_WORLD will produce the instrumentation
output file on behalf of the application and instrumentation data for the exited ranks will not be
included.
Instrumentation data now includes some information on messages sent to other MPI worlds
formed via MPI_Comm_accept(), MPI_Comm_connect() or MPI_Comm_join(). All off-world
message data is accounted together using the designation "offw" regardless of which off-world
rank was involved in the communication.
1.2.7.15 New environment variable MPI_ERROR_LEVEL
The new environment variable MPI_ERROR_LEVEL controls diagnostic output and abnormal
exit processing for application debugging where
0 Standard rank label text and abnormal exit processing. (Default)
1 Adds hostname and process id to rank label.
2 Adds hostname and process id to rank label. Also attempts to generate core file on abnormal
exit.
1.2.7.16 New flag -dbgspin
The mpirun command supports the new command line flag -dbgspin. This flag causes each
rank of the MPI application to spin in MPI_INIT(), allowing time for the user to log in to each
node running the MPI application and attach a debugger to each process. Setting the global
variable mpi_debug_cont to a non-zero value in the debugger will allow that process to continue.
This is similar to the debugging methods described in the mpidebug(1) manpage, except that
-dbgspin requires the user to launch and attach the debuggers manually.
1.2.7.17 Improved message rate
HP-MPI 2.2.7 has implemented message coalescing to improve the message rate of small messages.
When using many similar MPI_Isend() calls followed by an MPI_Waitall(), HP-MPI will
pack multiple Isend buffers together before sending the message. The receiver will automatically
unpack the buffers back into individual messages.
18 Information About This Release