Skip to content
This repository has been archived by the owner on Sep 30, 2022. It is now read-only.

fcoll/dynamic_gen2: add new component #987

Merged
merged 1 commit into from
Sep 14, 2016
Merged

fcoll/dynamic_gen2: add new component #987

merged 1 commit into from
Sep 14, 2016

Conversation

edgargabriel
Copy link
Member

this component is supposed to fix the performance problems of ompio on Lustre file systems.

bot:milestone:v2.0.1
bot:label:bug

@ompiteam-bot ompiteam-bot added this to the v2.0.1 milestone Feb 25, 2016
@edgargabriel edgargabriel changed the title fcoll/dynamioc_gen2: add new component fcoll/dynamic_gen2: add new component Feb 25, 2016
@mellanox-github
Copy link

Test PASSed.
See http://bgate.mellanox.com/jenkins/job/gh-ompi-release-pr/1373/ for details.

@jsquyres
Copy link
Member

bot:retest

@edgargabriel
Copy link
Member Author

@hppritcha this component is specifically designed for lustre file systems, to the best of my knowledge you are the only person having access to a lustre file system. Would you mind reviewing the new component?

@jsquyres
Copy link
Member

bot:retest

@mellanox-github
Copy link

Test PASSed.
See http://bgate.mellanox.com/jenkins/job/gh-ompi-release-pr/1917/ for details.

@hppritcha hppritcha modified the milestones: v2.1.0, v2.0.1 Jul 18, 2016
@hppritcha
Copy link
Member

RMs decided since this is really a new feature, even if fixing a performance problem, we'll move to 2.1 milestone.

@artpol84
Copy link
Contributor

artpol84 commented Aug 7, 2016

bot:mellanox:retest

@mellanox-github
Copy link

Test PASSed.
See http://bgate.mellanox.com/jenkins/job/gh-ompi-release-pr/2063/ for details.

@jsquyres
Copy link
Member

@edgargabriel This was originally committed 7 months ago. Is it still good / up-to-date?

@edgargabriel
Copy link
Member Author

yes, I just tested it 3 weeks back. I will do a rebase once we get close to merging it.

@jsquyres
Copy link
Member

@edgargabriel We're pretty close to merging. If you don't mind rebasing, that would be great. Thank you!

@edgargabriel
fcoll/dynamioc_gen2: add new component
bdd70ec 
this component is supposed to fix the performance problems of ompio on Lustre file systems.
@edgargabriel
Copy link
Member Author

ok, just did that.

@mellanox-github
Copy link

Test PASSed.
See http://bgate.mellanox.com/jenkins/job/gh-ompi-release-pr/2171/ for details.

@hppritcha
Copy link
Member

As long as I export OMPI_MCA_io=ompio the usual filetest passes with this PR on my lustre file system. Otherwise, if it uses romio (which seems to be what it does by default) there is a segfault in romio:

nid00111:04345] [nid00111:04346] [ 0] /lib64/libpthread.so.0(+0xf850)[0x7ffff7853850]
[nid00111:04346] [ 1] /cray/css/users/n17276/ompi_release_install/lib/openmpi/mca_io_romio314.so(ADIOI_Flatten+0x448)[0x7fffa6e7ae97]
[nid00111:04346] [ 2] [nid00111:04347] [ 0] /lib64/libpthread.so.0(+0xf850)[0x7ffff7853850]
[nid00111:04347] [ 1] /cray/css/users/n17276/ompi_release_install/lib/openmpi/mca_io_romio314.so(ADIOI_Flatten+0x448)[0x7fffa6e7ae97]
[nid00111:04347] [ 2] /cray/css/users/n17276/ompi_release_install/lib/openmpi/mca_io_romio314.so(ADIOI_Flatten_datatype+0x17f)[0x7fffa6e7aa3b]
[nid00111:04347] [ 3] [nid00111:04348] [ 0] /lib64/libpthread.so.0(+0xf850)[0x7ffff7853850]
[nid00111:04348] [ 1] /cray/css/users/n17276/ompi_release_install/lib/openmpi/mca_io_romio314.so(ADIOI_Flatten+0x448)[0x7fffa6e7ae97]
[nid00111:04348] [ 2] /cray/css/users/n17276/ompi_release_install/lib/openmpi/mca_io_romio314.so(ADIOI_Flatten_datatype+0x17f)[0x7fffa6e7aa3b]
[nid00111:04348] [ 3] [nid00111:04349] [ 0] /lib64/libpthread.so.0(+0xf850)[0x7ffff7853850]
[nid00111:04349] [ 1] /cray/css/users/n17276/ompi_release_install/lib/openmpi/mca_io_romio314.so(ADIOI_Flatten+0x448)[0x7fffa6e7ae97]
[nid00111:04349] [ 2] /cray/css/users/n17276/ompi_release_install/lib/openmpi/mca_io_romio314.so(ADIOI_Flatten_datatype+0x17f)[0x7fffa6e7aa3b]
[nid00111:04349] [ 3] /cray/css/users/n17276/ompi_release_install/lib/openmpi/mca_io_romio314.so(ADIO_Set_view+0x1b1)[0x7fffa6e70df5]
[nid00111:04349] [ 4] /cray/css/users/n17276/ompi_release_install/lib/openmpi/mca_io_romio314.so(ADIOI_Flatten_datatype+0x17f)[0x7fffa6e7aa3b]
[nid00111:04346] [ 3] /cray/css/users/n17276/ompi_release_install/lib/openmpi/mca_io_romio314.so(ADIO_Set_view+0x1b1)[0x7fffa6e70df5]
[nid00111:04346] [ 4] /cray/css/users/n17276/ompi_release_install/lib/openmpi/mca_io_romio314.so(mca_io_romio314_dist_MPI_File_set_view+0x5e0)[0x7fffa6e52bd4]
[nid00111:04346] [ 5] /cray/css/users/n17276/ompi_release_install/lib/openmpi/mca_io_romio314.so(mca_io_romio314_file_set_view+0x58)[0x7fffa6e4b35e]
[nid00111:04346] [ 6] /cray/css/users/n17276/ompi_release_install/lib/openmpi/mca_io_romio314.so(ADIO_Set_view+0x1b1)[0x7fffa6e70df5]
[nid00111:04347] [ 4] /cray/css/users/n17276/ompi_release_install/lib/openmpi/mca_io_romio314.so(mca_io_romio314_dist_MPI_File_set_view+0x5e0)[0x7fffa6e52bd4]
[nid00111:04347] [ 5] /cray/css/users/n17276/ompi_release_install/lib/openmpi/mca_io_romio314.so(mca_io_romio314_file_set_view+0x58)[0x7fffa6e4b35e]
[nid00111:04347] [ 6] /cray/css/users/n17276/ompi_release_install/lib/openmpi/mca_io_romio314.so(ADIO_Set_view+0x1b1)[0x7fffa6e70df5]
[nid00111:04348] [ 4] /cray/css/users/n17276/ompi_release_install/lib/openmpi/mca_io_romio314.so(mca_io_romio314_dist_MPI_File_set_view+0x5e0)[0x7fffa6e52bd4]
[nid00111:04348] [ 5] /cray/css/users/n17276/ompi_release_install/lib/openmpi/mca_io_romio314.so(mca_io_romio314_file_set_view+0x58)[0x7fffa6e4b35e]
[nid00111:04348] [ 6] /cray/css/users/n17276/ompi_release_install/lib/openmpi/mca_io_romio314.so(mca_io_romio314_dist_MPI_File_set_view+0x5e0)[0x7fffa6e52bd4]
[nid00111:04349] [ 5] /cray/css/users/n17276/ompi_release_install/lib/openmpi/mca_io_romio314.so(mca_io_romio314_file_set_view+0x58)[0x7fffa6e4b35e]
[nid00111:04349] [ 6] /cray/css/users/n17276/ompi_release_install/lib/libmpi.so.20(PMPI_File_set_view+0x1c2)[0x7ffff7b22a40]
[nid00111:04346] [ 7] /lus/dal/n17276/./filetest[0x403839]
[nid00111:04346] [ 8] /lus/dal/n17276/./filetest[0x4025aa]
[nid00111:04346] [ 9] /cray/css/users/n17276/ompi_release_install/lib/libmpi.so.20(PMPI_File_set_view+0x1c2)[0x7ffff7b22a40]
[nid00111:04347] [ 7] /lus/dal/n17276/./filetest[0x403839]
[nid00111:04347] [ 8] /lus/dal/n17276/./filetest[0x4025aa]
[nid00111:04347] [ 9] /cray/css/users/n17276/ompi_release_install/lib/libmpi.so.20(PMPI_File_set_view+0x1c2)[0x7ffff7b22a40]
[nid00111:04348] [ 7] /lus/dal/n17276/./filetest[0x403839]
[nid00111:04348] [ 8] /lus/dal/n17276/./filetest[0x4025aa]
[nid00111:04348] [ 9] /cray/css/users/n17276/ompi_release_install/lib/libmpi.so.20(PMPI_File_set_view+0x1c2)[0x7ffff7b22a40]
[nid00111:04349] [ 7] /lus/dal/n17276/./filetest[0x403839]
[nid00111:04349] [ 8] /lus/dal/n17276/./filetest[0x4025aa]
[nid00111:04349] [ 9] /lib64/libc.so.6(__libc_start_main+0xe6)[0x7ffff74e6c36]
[nid00111:04346] [10] /lus/dal/n17276/./filetest[0x4023a9]

since the purpose of this PR is for ompi and has nothing to do with romio, I'm good with this.

👍

@hppritcha
Copy link
Member

@jsquyres ready to merge

@jsquyres jsquyres merged commit 5220bcb into open-mpi:v2.x Sep 14, 2016
@mellanox-github
Copy link

Test FAILed.
See http://bgate.mellanox.com/jenkins/job/gh-ompi-release-pr/2188/ for details.

@edgargabriel
Copy link
Member Author

? Where does this failed Jenkins test suddenly come from?

@artpol84
Copy link
Contributor

artpol84 commented Oct 1, 2016

Indeed this is strange. I'll try to check.

But you can safely ignore it as this repo is empty now and failure is expected.

@mellanox-github
Copy link

Test FAILed.
See http://bgate.mellanox.com/jenkins/job/gh-ompi-release-pr/2189/ for details.

Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Reviewers
No reviews
Assignees

@hppritcha hppritcha

Labels
blocker bug enhancement reviewed RM-Approved
Projects
None yet
Milestone
v2.1.0
Development

Successfully merging this pull request may close these issues.
6 participants
@edgargabriel @mellanox-github @jsquyres @hppritcha @artpol84 @ompiteam-bot