Unable to run more than 32 processes.
Issue
Problem running more than 32 processes of a particular code. It works if we set it to a lower number. We are running on OpenMPI.
libibverbs: Warning: RLIMIT_MEMLOCK is 32768 bytes.
This will severely limit memory registrations.
libibverbs: Warning: RLIMIT_MEMLOCK is 32768 bytes.
This will severely limit memory registrations.
[compute-00-00.pcm:18084] 31 more processes have sent help message help-mpi-btl-base.txt / btl:no-nics
[compute-00-00.pcm:18084] Set MCA parameter "orte_base_help_aggregate" to 0 to see all help / error messages
[compute-00-03:16479] *** Process received signal ***
[compute-00-03:16479] Signal: Segmentation fault (11)
[compute-00-03:16479] Signal code: Address not mapped (1)
[compute-00-03:16479] Failing at address: 0x2aaab08e4280
[compute-00-03:16479] [ 0] /lib64/libpthread.so.0 [0x2ac2b4e50b10]
[compute-00-03:16479] [ 1] ../Code_src_July_29_2010_test/Para_FEBI_HOE_V1_test(totalmatrixluc_+0xf9a) [0x40d58a]
[compute-00-03:16479] [ 2] ../Code_src_July_29_2010_test/Para_FEBI_HOE_V1_test(febi_+0x3a1) [0x42e3d1]
[compute-00-03:16479] [ 3] ../Code_src_July_29_2010_test/Para_FEBI_HOE_V1_test(MAIN__+0x91) [0x404181]
[compute-00-03:16479] [ 4] ../Code_src_July_29_2010_test/Para_FEBI_HOE_V1_test(main+0xe) [0x42f21e]
[compute-00-03:16479] [ 5] /lib64/libc.so.6(__libc_start_main+0xf4) [0x2ac2b507b994]
[compute-00-03:16479] [ 6] ../Code_src_July_29_2010_test/Para_FEBI_HOE_V1_test [0x404039]
[compute-00-03:16479] *** End of error message ***
Environment
- Red Hat Cluster Suite
- OpenMPI.
Subscriber exclusive content
A Red Hat subscription provides unlimited access to our knowledgebase, tools, and much more.