search

hpc / ompi

Open MPI main development repository
Other
1 stars 2 forks source link

many failures when rebasing prototype on top of 856a2b7f #42

Closed hppritcha closed 3 years ago

hppritcha commented 3 years ago

For the record, many failures observed when using OFI MTL with the prototype after rebasing on top of master at 856a2b7f:

FAIL: allgather_gap_inter
FAIL: allreduce_nocommute_gap_inter
FAIL: reduce_scatter_nocommute_gap_inter
FAIL: reduce_scatter_block_nocommute_gap_inter
FAIL: iallgather_gap_inter
FAIL: iallgatherv_gap_inter
FAIL: iallreduce_nocommute_gap_inter
FAIL: igather_gap_inter
FAIL: igatherv_gap_inter
FAIL: ireduce_nocommute_gap_inter
FAIL: ireduce_scatter_nocommute_gap_inter
FAIL: ireduce_scatter_block_nocommute_gap_inter
FAIL: iscatter_gap_inter
FAIL: iscatterv_gap_inter
# XFAIL: 0
# FAIL:  14
FAIL: bcast_struct
FAIL: int_overflow
FAIL: op
FAIL: scatter_gap
FAIL: iallgather_gap
FAIL: iallreduce_nocommute_gap
FAIL: iallreduce_nocommute_gap_in_place
FAIL: ibcast_struct
FAIL: iexscan_nocommute_gap
FAIL: iexscan_nocommute_gap_in_place
FAIL: igather_gap
FAIL: ireduce_nocommute_gap
FAIL: ireduce_nocommute_gap_in_place
FAIL: ireduce_scatter_block_nocommute_gap
FAIL: ireduce_scatter_block_nocommute_gap_in_place
FAIL: ireduce_scatter_nocommute_gap
FAIL: ireduce_scatter_nocommute_gap_in_place
FAIL: iscan_nocommute_gap
FAIL: iscan_nocommute_gap_in_place
FAIL: iscatter_gap
FAIL: op_mpifh
FAIL: op_usempi
FAIL: op_usempif08
# XFAIL: 0
# FAIL:  23
# XFAIL: 0
# FAIL:  0
FAIL: aint
FAIL: aint_mpifh
FAIL: aint_usempi
FAIL: aint_usempif08
# XFAIL: 0
# FAIL:  4
# XFAIL: 0
# FAIL:  0
FAIL: loop_spawn
FAIL: no-disconnect
FAIL: intercomm_create
# XFAIL: 0
# FAIL:  3
FAIL: err
# XFAIL: 0
# FAIL:  1
# XFAIL: 0
# FAIL:  0
# XFAIL: 0
# FAIL:  0
FAIL: 60_info_env
# XFAIL: 0
# FAIL:  1
FAIL: c_fence_put_1
FAIL: c_put
FAIL: c_put_big
FAIL: c_accumulate
FAIL: c_accumulate_atomic
FAIL: c_accumulate_ddt1
FAIL: c_accumulate_ddt2
FAIL: c_accumulate_ddt3
FAIL: c_accumulate_ddt4
FAIL: c_get
FAIL: c_get_big
FAIL: c_transpose1
FAIL: c_transpose2
FAIL: c_transpose3
FAIL: c_transpose5
FAIL: c_transpose7
FAIL: c_strided_acc_indexed
FAIL: c_strided_acc_subarray
FAIL: c_strided_acc_onelock
FAIL: c_strided_get_indexed
FAIL: c_strided_putget_indexed
FAIL: c_post_start
FAIL: pp_1sided
FAIL: 1sided
FAIL: mt_1sided
FAIL: small_1sided
FAIL: multiple_locks2
FAIL: c_fetch_and_op
FAIL: c_flush
FAIL: c_get_accumulate
FAIL: c_get_accumulate_strided
FAIL: c_get_accumulate_ddt1
FAIL: c_get_accumulate_ddt2
FAIL: c_reqops
FAIL: c_strided_getacc_indexed
FAIL: halo_1sided_put_alloc_mem
FAIL: compare_and_swap
FAIL: ring_1sided_mpifh
FAIL: ring_1sided_sync_reg_mpifh
FAIL: halo_1sided_put_alloc_mem_usempi
FAIL: ring_1sided_usempi
FAIL: ring_1sided_sync_reg_usempi
FAIL: halo_1sided_put_alloc_mem_usempif08
FAIL: ring_1sided_usempif08
# XFAIL: 0
# FAIL:  44
FAIL: sendrecv_big
# XFAIL: 0
# FAIL:  1
# XFAIL: 0
# FAIL:  0
FAIL: random_seed
FAIL: attr-error-code
# XFAIL: 0
# FAIL:  2
# XFAIL: 0
# FAIL:  0

Same failures are observed with master at that sha, so proceeding boldy...

hppritcha commented 3 years ago

These failures appear to be due to commit 58c577739c691fa38e5a7bdcde3271d9abbf8454 on upstream master.

hppritcha commented 3 years ago

closing as we've rebased again to pull in fixes for at least some of the failures listed here.