Hi Folks,
When I start the mpi process (with openmpi 4.1.4) I can see jobid
getting printed as
$ mpirun -np 4 --map-by core --display-map ./osu_alltoall -i 1 -x 0 -m 128:128
Data for JOB [30105,1] offset 0 Total slots allocated 32
<snip>
Data for node: lib-ssp-04 Num slots: 32 Max slots: 0 Num procs: 4
Process OMPI jobid: [30105,1] App: 0 Process rank: 0 Bound:
socket 0[core 0[hwt
0]]:[B/././././././././././././././.][./././././././././././././././.]
Process OMPI jobid: [30105,1] App: 0 Process rank: 1 Bound:
socket 0[core 1[hwt
0]]:[./B/./././././././././././././.][./././././././././././././././.]
<snip>
$ tree pid.4025369
pid.4025369
___ 0
___ ___ 0
___ ___ debugger_attach_fifo
___ 1
___ ___ 0
___ ___ 1
___ ___ 2
___ ___ 3
___ contact.txt
___ hwloc.sm
___ pmix_dstor_ds12_4025369
___ ___ dstore_sm.lock
___ ___ initial-pmix_shared-segment-0
___ pmix_dstor_ds21_4025369
___ initial-pmix_shared-segment-0
___ smdataseg-1972961281-0 ------> What is this number 1972961281?
___ smlockseg-1972961281
___ smseg-1972961281-0
9 directories, 9 files
$ ps aux | grep mpirun
arun 4025369 0.0 0.0 309944 18800 pts/3 Sl+ 10:15 0:00
mpirun -np 4 --map-by core --display-map ./osu_alltoall -i 1 -x 0 -m
128:128 ---> Pid is 4025369
Can someone please explain
a) jobid: [30105,1], how jobid is determined?
b) how 1972961281 is derived? (smdataseg-1972961281-0)
c) What is the XY meaning in dsXY_4025369 ? ( pmix_dstor_ds12_4025369,
pmix_dstor_ds21_4025369)
--Arun