[Beowulf] RE: Compare and contrast MPI implementations
Mark Hahn
hahn at physics.mcmaster.ca
Tue Dec 20 13:27:15 PST 2005
> > rgb at lilith|B:1001>ssh lucifer "FROG='tasty';echo \$FROG"
> > tasty
ssh cat69 env FROG=tasty printenv
> > Or shell scripts. Or conditionals in your shell init files (or sshrc)
> > such as (pseudocode):
so far, I've avoided the modules approach - instead, I have a perl script
(/opt/compile/bin/compile, with symlinks so it appears as cc/f77/f90/c++).
it's like a super mpicc, except it handles different languages, compilers,
MPIs, and clusters. new users generally use it like
f90 *.f90 -lmpi -o prog
but it works perfectly well in a makefile. to get PGI:
f90 -pgi *.f90 -lmpi -o prog
the script chooses compiler-specific opt flags, unless the user provides them.
(this script was originally born of the need to log how often any user
compiled, so that we could provide "feedback" to Pathscale on their licensing
scheme. it's rather disappointing to find out how often users compile code
with something like "gcc *.c -o whatever" and expect it to work well without
optimization flags...
More information about the Beowulf
mailing list