Hello, I have troubles with latest trunk if I use PMI1.
For example, if I use 2 nodes the application hangs. See backtraces from
both nodes below. From them I can see that second (non launching) node
hangs in bcol component selection. Here is the default setting of
bcol_base_string parameter:
bcol_base_string="basesmuma,basesmuma,iboffload,ptpcoll,ugni"
according to ompi_info. I don't know if it is correct that basesmuma is
duplicated or not.
Experiments with this parameter showed that it directly influences the bug:
export OMPI_MCA_bcol_base_string="" # [SEGFAULT]
export OMPI_MCA_bcol_base_string="ptpcoll" # [OK]
export OMPI_MCA_bcol_base_string="basesmuma,ptpcoll" # [OK]
export OMPI_MCA_bcol_base_string="basesmuma,ptpcoll,iboffload" # [OK]
export OMPI_MCA_bcol_base_string="basesmuma,ptpcoll,iboffload,ugni" # [OK]
export
OMPI_MCA_bcol_base_string="basesmuma,basesmuma,ptpcoll,iboffload,ugni" #
[HANG]
export OMPI_MCA_bcol_base_string="basesmuma,basesmuma,iboffload,ptpcoll" #
[HANG]
export OMPI_MCA_bcol_base_string="basesmuma,basesmuma,iboffload" # [OK]
export OMPI_MCA_bcol_base_string="basesmuma,basesmuma,iboffload,ugni" # [OK]
export OMPI_MCA_bcol_base_string="basesmuma,basesmuma,ptpcoll" # [HANG]
export OMPI_MCA_bcol_base_string="ptpcoll,basesmuma" # [OK]
export OMPI_MCA_bcol_base_string="ptpcoll,basesmuma,basesmuma" # [HANG]
I can provide other information if nessesary.
cn1:
(gdb) bt
0 0x00007fdebd30ac6d in poll () from /lib/x86_64-linux-gnu/libc.so.6
1 0x00007fdebcca64e0 in poll_dispatch (base=0x1d466b0, tv=0x7fff71aab880)
at poll.c:165
2 0x00007fdebcc9b041 in opal_libevent2021_event_base_loop (base=0x1d466b0,
flags=2) at event.c:1631
3 0x00007fdebcc35891 in opal_progress () at runtime/opal_progress.c:169
4 0x00007fdeb32f78cb in opal_condition_wait (c=0x7fdebdb51bc0
<ompi_request_cond>, m=0x7fdebdb51cc0 <ompi_request_lock>) at
../../../../opal/threads/condition.h:78
5 0x00007fdeb32f79b8 in ompi_request_wait_completion (req=0x7fff71aab920)
at ../../../../ompi/request/request.h:381
6 0x00007fdeb32f84b8 in mca_pml_ob1_recv (addr=0x7fff71aabd80, count=1,
datatype=0x6026c0 <ompi_mpi_int>, src=1, tag=0, comm=0x6020a0
<ompi_mpi_comm_world>,
status=0x7fff71aabd90) at pml_ob1_irecv.c:109
7 0x00007fdebd88f54d in PMPI_Recv (buf=0x7fff71aabd80, count=1,
type=0x6026c0 <ompi_mpi_int>, source=1, tag=0, comm=0x6020a0
<ompi_mpi_comm_world>,
status=0x7fff71aabd90) at precv.c:78
8 0x0000000000400c44 in main (argc=1, argv=0x7fff71aabe98) at hellompi.c:33
cn2:
(gdb) bt
0 0x00007fa65aa78c6d in poll () from /lib/x86_64-linux-gnu/libc.so.6
1 0x00007fa65a4144e0 in poll_dispatch (base=0x20e96b0, tv=0x7fff46f44a80)
at poll.c:165
2 0x00007fa65a409041 in opal_libevent2021_event_base_loop (base=0x20e96b0,
flags=2) at event.c:1631
3 0x00007fa65a3a3891 in opal_progress () at runtime/opal_progress.c:169
4 0x00007fa65afbbc25 in opal_condition_wait (c=0x7fa65b2bfbc0
<ompi_request_cond>, m=0x7fa65b2bfcc0 <ompi_request_lock>) at
../opal/threads/condition.h:78
5 0x00007fa65afbc1b5 in ompi_request_default_wait_all (count=2,
requests=0x7fff46f44c70, statuses=0x0) at request/req_wait.c:287
6 0x00007fa65afc7906 in comm_allgather_pml (src_buf=0x7fff46f44da0,
dest_buf=0x233dac0, count=288, dtype=0x7fa65b29fee0 <ompi_mpi_char>,
my_rank_in_group=1,
n_peers=2, ranks_in_comm=0x210a760, comm=0x6020a0
<ompi_mpi_comm_world>) at patterns/comm/allgather.c:250
7 0x00007fa64f14ba08 in bcol_basesmuma_smcm_allgather_connection
(sm_bcol_module=0x7fa64e64d010, module=0x232c800,
peer_list=0x7fa64f3513e8 <mca_bcol_basesmuma_component+456>,
back_files=0x7fa64eae2690, comm=0x6020a0 <ompi_mpi_comm_world>, input=...,
base_fname=0x7fa64f14ca8c "sm_ctl_mem_", map_all=false) at
bcol_basesmuma_smcm.c:205
8 0x00007fa64f146525 in base_bcol_basesmuma_setup_ctl
(sm_bcol_module=0x7fa64e64d010, cs=0x7fa64f351220
<mca_bcol_basesmuma_component>) at bcol_basesmuma_setup.c:344
9 0x00007fa64f146cbb in base_bcol_basesmuma_setup_library_buffers
(sm_bcol_module=0x7fa64e64d010, cs=0x7fa64f351220
<mca_bcol_basesmuma_component>)
at bcol_basesmuma_setup.c:550
10 0x00007fa64f1418d0 in mca_bcol_basesmuma_comm_query (module=0x232c800,
num_modules=0x232e570) at bcol_basesmuma_module.c:532
11 0x00007fa64fd9e5f2 in mca_coll_ml_tree_hierarchy_discovery
(ml_module=0x232fbe0, topo=0x232fd98, n_hierarchies=3,
exclude_sbgp_name=0x0, include_sbgp_name=0x0)
at coll_ml_module.c:1964
12 0x00007fa64fd9f3a3 in mca_coll_ml_fulltree_hierarchy_discovery
(ml_module=0x232fbe0, n_hierarchies=3) at coll_ml_module.c:2211
13 0x00007fa64fd9cbe4 in ml_discover_hierarchy (ml_module=0x232fbe0) at
coll_ml_module.c:1518
14 0x00007fa64fda164f in mca_coll_ml_comm_query (comm=0x6020a0
<ompi_mpi_comm_world>, priority=0x7fff46f45358) at coll_ml_module.c:2970
15 0x00007fa65b02f6aa in query_2_0_0 (component=0x7fa64fffe4e0
<mca_coll_ml_component>, comm=0x6020a0 <ompi_mpi_comm_world>,
priority=0x7fff46f45358,
module=0x7fff46f45390) at base/coll_base_comm_select.c:374
16 0x00007fa65b02f66e in query (component=0x7fa64fffe4e0
<mca_coll_ml_component>, comm=0x6020a0 <ompi_mpi_comm_world>,
priority=0x7fff46f45358, module=0x7fff46f45390)
at base/coll_base_comm_select.c:357
17 0x00007fa65b02f581 in check_one_component (comm=0x6020a0
<ompi_mpi_comm_world>, component=0x7fa64fffe4e0 <mca_coll_ml_component>,
module=0x7fff46f45390)
at base/coll_base_comm_select.c:319
18 0x00007fa65b02f3c7 in check_components (components=0x7fa65b2a9530
<ompi_coll_base_framework+80>, comm=0x6020a0 <ompi_mpi_comm_world>)
at base/coll_base_comm_select.c:283
19 0x00007fa65b027d45 in mca_coll_base_comm_select (comm=0x6020a0
<ompi_mpi_comm_world>) at base/coll_base_comm_select.c:119
20 0x00007fa65afbdb2c in ompi_mpi_init (argc=1, argv=0x7fff46f45a78,
requested=0, provided=0x7fff46f4590c) at runtime/ompi_mpi_init.c:858
21 0x00007fa65aff20ef in PMPI_Init (argc=0x7fff46f4594c,
argv=0x7fff46f45940) at pinit.c:84
22 0x0000000000400b66 in main (argc=1, argv=0x7fff46f45a78) at hellompi.c:11
--
С Уважением, Поляков Артем Юрьевич
Best regards, Artem Y. Polyakov
