charm AT lists.siebelschool.illinois.edu
Subject: Charm++ parallel programming system
List archive
- From: vasudevan rengasamy <vas.renga AT gmail.com>
- To: charm AT cs.uiuc.edu
- Subject: [charm] Charm++ message priorities
- Date: Mon, 09 Jul 2012 19:36:59 -0000
- List-archive: <http://lists.cs.uiuc.edu/pipermail/charm>
- List-id: CHARM parallel programming system <charm.cs.uiuc.edu>
Sir,
In HybridAPI, how can I specify that the callback function (invoked after the CUDA kernel execution of a workRequest) is called with high priority so it gets processed immediately?
Can I achieve this by making the callback function an expedited entry method?
Is there a way to specify priority for callback messages?
Please let me know if you need any specific details.
Thanks,
R. Vasudevan,
Super Computation Education and Research Centre,
Indian Institute of Science.
- [charm] Charm++ message priorities, vasudevan rengasamy, 07/09/2012
- Re: [charm] Charm++ message priorities, vasudevan rengasamy, 07/17/2012
- Re: [charm] [ppl] Charm++ message priorities, Phil Miller, 07/17/2012
- Re: [charm] Charm++ message priorities, vasudevan rengasamy, 07/17/2012
Archive powered by MHonArc 2.6.16.