On Sun, 23 Apr 2006 10:40:12 -0500 (CDT) Maura Edeweiss Monville <memonvil@artsci.wustl.edu> wrote:
But can anyone explain WHY the same Monte Carlo installation (MC routine + CERN library + Tcl/Tk files) is RUNNING in a research center in Japan where my supervisor has installed on a x86_64 Linux/Red-Hat clusters. He installed by recompiling everything on that cluster with the -m32 option, just like on my laptop. We remotely connected from here to the Japanese center. We compared the Makefiles etc .. They coincide. But there it works. Whereas for me it doesn't.
Someone asked me about the Tcl/Tk source. The answer is "yes". The procedure that installes and generates the Tcl/Tk ".a" and ".h" files does that starting from Tcl/Tk 8.3 source.
This is Linux mistery, isn't it ? You just gave me a clue. Check the versions of GCC. What version of Red Hat is he using and what version of gcc. RHEL 3.0 (2.4 kernel) or RHEL 4.0 (2.6 kernel) There were a number of changes made in GCC. I forget exactly what version of GCC is installed on RHEL, but it is either 3.3.3 or 3.4.5 or possible even 3.2.x.
You can legitimately have multiple versions of gcc on the system. -- Jerry Feldman <gaf@blu.org> Boston Linux and Unix user group http://www.blu.org PGP key id:C5061EA9 PGP Key fingerprint:053C 73EC 3AC1 5C44 3E14 9245 FB00 3ED5 C506 1EA9