charm AT lists.siebelschool.illinois.edu
Subject: Charm++ parallel programming system
List archive
- From: Gengbin Zheng <gzheng AT illinois.edu>
- To: Blesson Varghese <b.varghese AT student.reading.ac.uk>
- Cc: charm AT cs.uiuc.edu
- Subject: Re: [charm] Mapping Scheme
- Date: Mon, 23 Nov 2009 11:08:59 -0600
- List-archive: <http://lists.cs.uiuc.edu/pipermail/charm>
- List-id: CHARM parallel programming system <charm.cs.uiuc.edu>
the following section describe how to write a map object:
http://charm.cs.uiuc.edu/manuals/html/charm++/3_8.html#SECTION00038400000000000000
When you have defined your own mapping, you need to publish the new mapping in charm runtime. Look at the function:
TCHARM_Build_threads()
in charm/src/libs/ck-libs/tcharm, where it parses "mapping" command line option to create the corresponding mapping object.
Gengbin
Blesson Varghese wrote:
Hi,
I am just referring to the AMPI Manual Version 1.0, Page 15 regarding
Mapping Schemes. I would like to define my own mapping scheme of virtual
processors to physical processors as a run time option.
Would you please advise on how it could be done?
Many thanks and best regards,
Blesson.
_______________________________________________
charm mailing list
charm AT cs.uiuc.edu
http://lists.cs.uiuc.edu/mailman/listinfo/charm
- [charm] Mapping Scheme, Blesson Varghese, 11/19/2009
- <Possible follow-up(s)>
- [charm] Mapping Scheme, Blesson Varghese, 11/23/2009
- Re: [charm] Mapping Scheme, Gengbin Zheng, 11/23/2009
Archive powered by MHonArc 2.6.16.