openmpi/contrib/scaling/Makefile

PROGS = orte_no_op mpi_no_op mpi_memprobe

all: $(PROGS)

CFLAGS = -O

orte_no_op: orte_no_op.c
	ortecc -o orte_no_op orte_no_op.c

mpi_no_op: mpi_no_op.c
	mpicc -o mpi_no_op mpi_no_op.c

mpi_memprobe: mpi_memprobe.c
	mpicc -o mpi_memprobe mpi_memprobe.c -lopen-pal -lopen-rte

clean:
	rm -f $(PROGS) *~
Remove stale global variables Revamp the event notification integration to rely on the PMIx event chaining and remove the duplicate chaining in OPAL. This ensures we get system-level events that target non-default handlers. Restore the hostname entries for MPI-level error messages, but provide an MCA param (orte_hostname_cutoff) to remove them for large clusters where the memory footprint is problematic. Set the default at 1000 nodes in the job (not the allocation). Begin first cut at memory profiler Some minor cleanups of memprobe Signed-off-by: Ralph Castain <rhc@open-mpi.org> 2016-12-31 07:37:24 -08:00			`PROGS = orte_no_op mpi_no_op mpi_memprobe`
Add scaling tests and script This commit was SVN r26509. 2012-05-29 15:21:44 +00:00
			`all: $(PROGS)`

Use preconnect as a better test of startup scaling than barrier This commit was SVN r26530. 2012-06-01 02:35:15 +00:00			`CFLAGS = -O`
Update the scaling test script 2015-11-06 21:51:40 -08:00
Complete the memprobe support. This provides a new scaling tool called "mpi_memprobe" that samples the memory footprint of the local daemon and the client procs, and then reports the results. The output contains the footprint of the daemon on each node, plus the average footprint of the client procs on that node. Samples are taken after MPI_Init, and then again after MPI_Barrier. This allows the user to see memory consumption caused by add_procs, as well as any modex contribution from forming connections if pmix_base_async_modex is given. Using the probe simply involves executing it via mpirun, with however many copies you want per node. Example: $ mpirun -npernode 2 ./mpi_memprobe Sampling memory usage after MPI_Init Data for node rhc001 Daemon: 12.483398 Client: 6.514648 Data for node rhc002 Daemon: 11.865234 Client: 4.643555 Sampling memory usage after MPI_Barrier Data for node rhc001 Daemon: 12.520508 Client: 6.576660 Data for node rhc002 Daemon: 11.879883 Client: 4.703125 Note that the client value on node rhc001 is larger - this is where rank=0 is housed, and apparently it gets a larger footprint for some reason. Signed-off-by: Ralph Castain <rhc@open-mpi.org> 2017-01-05 09:46:05 -08:00			`orte_no_op: orte_no_op.c`
Update the scaling test script 2015-11-06 21:51:40 -08:00			`ortecc -o orte_no_op orte_no_op.c`

Complete the memprobe support. This provides a new scaling tool called "mpi_memprobe" that samples the memory footprint of the local daemon and the client procs, and then reports the results. The output contains the footprint of the daemon on each node, plus the average footprint of the client procs on that node. Samples are taken after MPI_Init, and then again after MPI_Barrier. This allows the user to see memory consumption caused by add_procs, as well as any modex contribution from forming connections if pmix_base_async_modex is given. Using the probe simply involves executing it via mpirun, with however many copies you want per node. Example: $ mpirun -npernode 2 ./mpi_memprobe Sampling memory usage after MPI_Init Data for node rhc001 Daemon: 12.483398 Client: 6.514648 Data for node rhc002 Daemon: 11.865234 Client: 4.643555 Sampling memory usage after MPI_Barrier Data for node rhc001 Daemon: 12.520508 Client: 6.576660 Data for node rhc002 Daemon: 11.879883 Client: 4.703125 Note that the client value on node rhc001 is larger - this is where rank=0 is housed, and apparently it gets a larger footprint for some reason. Signed-off-by: Ralph Castain <rhc@open-mpi.org> 2017-01-05 09:46:05 -08:00			`mpi_no_op: mpi_no_op.c`
Update the scaling test script 2015-11-06 21:51:40 -08:00			`mpicc -o mpi_no_op mpi_no_op.c`
Add scaling tests and script This commit was SVN r26509. 2012-05-29 15:21:44 +00:00
Complete the memprobe support. This provides a new scaling tool called "mpi_memprobe" that samples the memory footprint of the local daemon and the client procs, and then reports the results. The output contains the footprint of the daemon on each node, plus the average footprint of the client procs on that node. Samples are taken after MPI_Init, and then again after MPI_Barrier. This allows the user to see memory consumption caused by add_procs, as well as any modex contribution from forming connections if pmix_base_async_modex is given. Using the probe simply involves executing it via mpirun, with however many copies you want per node. Example: $ mpirun -npernode 2 ./mpi_memprobe Sampling memory usage after MPI_Init Data for node rhc001 Daemon: 12.483398 Client: 6.514648 Data for node rhc002 Daemon: 11.865234 Client: 4.643555 Sampling memory usage after MPI_Barrier Data for node rhc001 Daemon: 12.520508 Client: 6.576660 Data for node rhc002 Daemon: 11.879883 Client: 4.703125 Note that the client value on node rhc001 is larger - this is where rank=0 is housed, and apparently it gets a larger footprint for some reason. Signed-off-by: Ralph Castain <rhc@open-mpi.org> 2017-01-05 09:46:05 -08:00			`mpi_memprobe: mpi_memprobe.c`
			`mpicc -o mpi_memprobe mpi_memprobe.c -lopen-pal -lopen-rte`
Remove stale global variables Revamp the event notification integration to rely on the PMIx event chaining and remove the duplicate chaining in OPAL. This ensures we get system-level events that target non-default handlers. Restore the hostname entries for MPI-level error messages, but provide an MCA param (orte_hostname_cutoff) to remove them for large clusters where the memory footprint is problematic. Set the default at 1000 nodes in the job (not the allocation). Begin first cut at memory profiler Some minor cleanups of memprobe Signed-off-by: Ralph Castain <rhc@open-mpi.org> 2016-12-31 07:37:24 -08:00
Add scaling tests and script This commit was SVN r26509. 2012-05-29 15:21:44 +00:00			`clean:`
			`rm -f $(PROGS) *~`