From: Peter Freddolino (petefred_at_ks.uiuc.edu)
Date: Mon Jul 02 2007 - 10:32:56 CDT
Hi Justin,
in general you're probably better off addressing charm++ issues to the
http://lists.cs.uiuc.edu/mailman/listinfo/charm since there isn't much
overlap in readership between the two lists.
In this case, however, is there any reason you're not just using the
precompiled bluegene binary? It makes use of an unreleased (by IBM)
communication layer, and achieves much better scaling than MPI.
Best,
Peter
Jianhui Tian wrote:
> Hi,
>
> I am compiling Charm++, and got the following error.
> ########################################################
> ../bin/charmc -host xi-main.C
> ../bin/charmc: line 160: xlC: command not found
> Fatal Error by charmc in directory
> /gpfs1/angel/NAMD/NAMD_2.6_Source/charm-5.9/mpi-bluegenel-xlc/tmp
> Command xlC -I../bin/../include -D__CHARMC__=1
> -I/bgl/BlueLight/ppcfloor//bglsys/include -c xi-main.C -o xi-main.o
> returned error code 127
> charmc exiting...
> gmake[1]: *** [xi-main.o] Error 1
> gmake[1]: Leaving directory
> `/gpfs1/angel/NAMD/NAMD_2.6_Source/charm-5.9/mpi-bluegenel-xlc/tmp'
> gmake: *** [headers] Error 2
> -------------------------------------------------
> Charm++ NOT BUILT.
> #######################################
> I checked CMK_CXX="/opt/ibmcmp/vacpp/bg/8.0/bin/blrts_xlC", blrts_xlC
> is there. Anybody has ang hint about what's wrong here? Thanks a lot.
>
> Justin
This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:44:54 CST