search

pmodels / mpich

Official MPICH Repository
http://www.mpich.org
Other
541 stars 280 forks source link

pamid: request-based rma operations #1984

Closed mpichbot closed 8 years ago

mpichbot commented 8 years ago

Originally by blocksom on 2013-12-05 11:01:06 -0600

Adding support for request-based rma operation in pamid.

This touches non-pamid code and needs to be code reviewed.

mpichbot commented 8 years ago

Originally by blocksom on 2013-12-05 11:14:30 -0600

I pushed the changes to the branch mpich-ibm/ticket-1984. Ready for review!

mpichbot commented 8 years ago

Originally by blocksom on 2013-12-17 13:16:26 -0600

Can someone at ANL take a look at this? I think the changes to the common code is fairly minor.

I'd like to get this change into the mpich/master branch soon (this week?) before developers start to check-out for the year.

Thanks!

mpichbot commented 8 years ago

Originally by blocksom on 2013-12-18 13:43:15 -0600

I pushed a change to master that causes this mpich-ibm/ticket-1984 branch to have conflicts when I did a rebase. I'm working to fix it up now and will update the ticket again shortly.

mpichbot commented 8 years ago

Originally by blocksom on 2013-12-19 11:07:54 -0600

I've deleted the mpich-ibm/ticket-1984 branch, rebased to code to include the recent pami win-get-acc changes from Sameer, and pushed the code back to the mpich-ibm/ticket-1984 branch.

I didn't rebase all the way to the HEAD of the mpich/master branch because of ticket #1991.

This code is again ready to review and integrate.

mpichbot commented 8 years ago

Originally by @raffenet on 2013-12-19 11:30:15 -0600

Mike, I pushed that branch to the review queue and it fails to build on our Jenkins slaves. You can see an example build result here:

https://jenkins.mpich.org/view/mpich-review/job/mpich-review/OS=ubuntu64/507/console

mpichbot commented 8 years ago

Originally by blocksom on 2013-12-19 11:43:45 -0600

Thanks, Ken .. we are looking into the compile problem now.

mpichbot commented 8 years ago

Originally by blocksom on 2013-12-19 16:05:01 -0600

I've push a fix from Haizhu into branch mpich-ibm/ticket-1984-new and have deleted the old branch.

Haizhu tested on PE, and I did a test compile on bgq and I tested on our ppc linux front-end nodes with the configure from the failing jenkins build:

configure --enable-xfail --enable-nemesis-dbg-localoddeven

This code is (again) ready to review and integrate. Thanks!

mpichbot commented 8 years ago

Originally by @raffenet on 2013-12-19 17:16:12 -0600

I pushed the new branch to review and tagged it for Pavan to look at. Jenkins looks to have built this one fine.

mpichbot commented 8 years ago

Originally by blocksom on 2013-12-20 15:31:33 -0600

We have a pamid-only commit that depends on this request-based ops code in this ticket. I've pushed this dependent commit to mpich-ibm/ticket-1984-haizhu-atomics.

Please push this commit to master when the 1984 commit ("PAMID support of request based rma functions") has been pushed.

Thanks!

mpichbot commented 8 years ago

Originally by balaji on 2013-12-31 12:59:22 -0600

I've merged a slight variant of mpich-ibm/ticket-1984-new to mpich/master in [d389a534] and mpich-ibm/ticket-1984-haizhu-atomics in [8c5ab383].

Resolving this ticket. Please reopen if I missed something.