Charm++ parallel programming system

[Phil Miller <mille121 AT illinois.edu>]

By default, messages are delivered in FIFO order on each PE, and execution proceeds accordingly. In Charm++ code, it's possible to set desired message priorities and queueing order, according to the manual
http://charm.cs.illinois.edu/manuals/html/charm++/10.html#SECTION02130000000000000000

This gives the runtime system hints on the order that the developer expects to produce the best performance, but doesn't offer any guarantee of execution order.

AMPI doesn't currently offer a way to directly prioritize its messages. Please let us know if this arises as a performance problem in an application you're working with.

On Tue, Nov 11, 2014 at 6:40 AM, Rafael Keller Tesser <rktesser AT inf.ufrgs.br> wrote:

Hello!

I'd like to know if there's any documentation on how the execution
Chares (or VPs) mapped to the same process are scheduled. More
specifically, I'm interested on how this works in AMPI, if there's any
difference from pure Charm++.

Regards,
Rafael Keller Tesser
_______________________________________________
charm mailing list
charm AT cs.uiuc.edu
http://lists.cs.uiuc.edu/mailman/listinfo/charm