Prevent forward events from overtaking cancel events.

[nikhil-jain]

Add compile time option to stop sending of events from a PE if a cancel event is being held for its corresponding forward event from being sent.

cmake option: PRIORITIZE_CANCEL_EVENTS default value: OFF

This is still no fool-proof because on the receive side, MPI_Testsome may succeed for events sent later. However, for small messages that ROSS usually sends (<1 KB), this is unlikely.

Change-Id: I2800a2a6d3537317c6182047ff0beeb3ba238037

@mmubarak - you may be interested in this.

If this merge represents a feature addition to ROSS, the following items must be completed before the branch will be merged:

• [ ] Document the feature on the blog (See the Contributing guide in the gh-pages branch). Include a link to your blog post in the Pull Request.
• [ ] One or more TravisCI tests should be created (and they should pass)
• [ ] Through the TravisCI tests, coverage should increase
• [ ] Test with CODES to ensure everything continues to work

[codecov-io]

Codecov Report

Merging #123 into master will not change coverage. The diff coverage is n/a.

@@           Coverage Diff           @@
##           master     #123   +/-   ##
=======================================
  Coverage   75.24%   75.24%           
=======================================
  Files          32       32           
  Lines        4039     4039           
  Branches      752      752           
=======================================
  Hits         3039     3039           
  Misses        973      973           
  Partials       27       27

Impacted Files	Coverage Δ
core/network-mpi.c	80.14% <ø> (ø)	:arrow_up:

---

Continue to review full report at Codecov.

Legend - Click here to learn more Δ = absolute <relative> (impact), ø = not affected, ? = missing data Powered by Codecov. Last update cb2377e...85c5f1c. Read the comment docs.

[nikhil-jain]

Thoughts? @carothersc

[Starman911]

Can you explain the process of your project, and what you need accomplished?

On Aug 14, 2017, at 3:02 PM, Nikhil Jain [email protected] wrote:

Thoughts? @carothersc https://github.com/carothersc — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/carothersc/ROSS/pull/123#issuecomment-322279162, or mute the thread https://github.com/notifications/unsubscribe-auth/AcyV1cfGfuOHCWJs_HbyJPVn14rPUpyjks5sYJmrgaJpZM4OnQ6u.