Skip to Content.
Sympa Menu

charm - Re: [charm] compilation problem

charm AT lists.cs.illinois.edu

Subject: Charm++ parallel programming system

List archive

Re: [charm] compilation problem


Chronological Thread 
  • From: Phil Miller <mille121 AT illinois.edu>
  • To: Sabrina Bergoch <sabrinabms AT gmail.com>
  • Cc: "charm AT cs.uiuc.edu" <charm AT cs.uiuc.edu>
  • Subject: Re: [charm] compilation problem
  • Date: Wed, 26 Nov 2014 13:24:56 -0600
  • List-archive: <http://lists.cs.uiuc.edu/pipermail/charm/>
  • List-id: CHARM parallel programming system <charm.cs.uiuc.edu>

Hi Sabrina,

We're actually in the process of changing our documentation to suggest that users simply compile their new load balancing strategies independently of the core system. This doesn't change anything about their performance or how to link them in - solely when and where they get built. If you subsequently want to suggest that we include the LB strategy you've written in our distribution, we'll take care of the various script updates and build system incorporation.

Here's the summary version of the way this would go:
1. charmc NameLB.ci
2. charmc -o libmoduleNameLB.a NameLB.C
2. charmc -o application_name foo.o bar.o baz.o -module NameLB

Hope this helps.

Phil

On Wed, Nov 26, 2014 at 1:18 PM, Sabrina Bergoch <sabrinabms AT gmail.com> wrote:
Hi,

I'm trying do include a new load balancer on Charm++. 
What I did:
  • Put the file in the directory: ./charm-20141124/tmp
  • Run: Makefile_lb.sh (for creating Make.lb)
  • Generate the balancer interface using "charmc NameLB.ci"
  • Run: make depends (*)
(*) In this step, I dropped in the trouble. The system informes to me: 
(after several checking dependencies, appears)
   "checking dependencies for ScotchLB.C
    checking dependencies for TeamLB.C
    TeamLB.C:13:19: fatal error: metis.h: No such file or directory
    compilation terminated.

    Compilation of 'TeamLB.C' failed, please fix it first!!!!"

The compilation fail because the system could not find the file "metis.h".
Actually, when I did a comment on this checking point the trouble repeats in other checking.

Does someone have one idea how can I solve this problem?

I'd appreciate any help.

Sabrina.

_______________________________________________
charm mailing list
charm AT cs.uiuc.edu
http://lists.cs.uiuc.edu/mailman/listinfo/charm





Archive powered by MHonArc 2.6.16.

Top of Page