[mpich-devel] Assertion failed in file src/mpid/ch3/channels/nemesis/src/ch3_progress.c at line 643: pkt->type >= 0 && pkt->type < MPIDI_CH3_PKT_END_ALL internal ABORT - process 1

[Balaji, Pavan]

Do you have more details on how mpich was built?

  -- Pavan

> On Dec 29, 2015, at 2:05 PM, Jeff Hammond <jeff.science at gmail.com> wrote:
> 
> I am using mpich_3.2b3-1ubuntu_amd64.deb within Travis CI for ARMCI-MPI testing.  I never saw this problem in Jenkins, so I can't tell if the problem is specific to the deb build, to Travis, or something else.
> 
> Travis page has all the details so I will not repeat them here.
> 
> Thanks,
> 
> Jeff
> 
> From https://travis-ci.org/jeffhammond/armci-mpi/jobs/99316268:
> 
>  Processing triggers for libc-bin ...
> ldconfig deferred processing now taking place
> +wget -q http://www.cebacad.net/files/mpich/ubuntu/mpich-3.2b3/mpich_3.2b3-1ubuntu_amd64.deb
> +sudo dpkg -i ./mpich_3.2b3-1ubuntu_amd64.deb
> Selecting previously unselected package mpich.
> (Reading database ... 124680 files and directories currently installed.)
> Unpacking mpich (from .../mpich_3.2b3-1ubuntu_amd64.deb) ...
> Setting up mpich (3.2b3-1ubuntu) ...
> Processing triggers for man-db ...
> 
> 
> FAIL: benchmarks/contiguous-bench (exit: 1)
> ===========================================
> ARMCI-MPI initialized with 2 processes, MPI v3.1
>   PROGRESS_THREAD        = DISABLED
>   ALLOC_SHM used         = TRUE
>   WINDOW type used       = CREATE
>   STRIDED_METHOD         = DIRECT
>   IOV_METHOD             = DIRECT
>   IOV_CHECKS             = FALSE
>   SHR_BUF_METHOD         = COPY
>   NONCOLLECTIVE_GROUPS   = FALSE
>   CACHE_RANK_TRANSLATION = TRUE
>   DEBUG_ALLOC            = FALSE
> Starting one-sided contiguous performance test with 2 processes
>    Trg. Rank    Xfer Size   Get (usec)   Put (usec)   Acc (usec)  Get (MiB/s)  Put (MiB/s)  Acc (MiB/s)
>            1            8        0.869        0.707        0.935        8.783       10.795        8.158
>            1           16        0.926        0.768        0.914       16.482       19.859       16.698
>            1           32        0.945        0.856        0.974       32.284       35.671       31.336
>            1           64        0.938        0.781        0.956       65.097       78.199       63.852
>            1          128        0.946        0.833        1.055      128.992      146.623      115.730
>            1          256        0.969        0.819        1.065      251.986      297.954      229.235
>            1          512        0.953        0.867        1.024      512.188      562.918      477.005
>            1         1024        0.995        0.873        1.139      981.813     1119.078      857.161
>            1         2048        1.010        0.850        1.044     1932.970     2297.619     1870.578
>            1         4096        1.071        1.172        1.494     3647.221     3334.105     2614.495
>            1         8192        1.634        1.721        2.068     4780.515     4539.290     3777.801
>            1        16384        2.989        3.325        4.856     5227.361     4698.842     3217.416
>            1        32768        5.321        7.480        9.465     5872.833     4177.594     3301.464
>            1        65536       13.295       14.356       20.259     4701.146     4353.478     3085.039
>            1       131072       23.574       31.835       45.756     5302.534     3926.447     2731.862
>            1       262144       44.921       70.181      102.028     5565.275     3562.231     2450.313
> Assertion failed in file src/mpid/ch3/channels/nemesis/src/ch3_progress.c at line 643: pkt->type >= 0 && pkt->type < MPIDI_CH3_PKT_END_ALL
> internal ABORT - process 1
> 
> 
> -- 
> Jeff Hammond
> jeff.science at gmail.com
> http://jeffhammond.github.io/
> _______________________________________________
> To manage subscription options or unsubscribe:
> https://lists.mpich.org/mailman/listinfo/devel