debian-beowulf appears to have very little activity. Perhaps I can get some help here?
---------- Forwarded Message ---------- Subject: before bug report: perhaps I can get help here Date: Mon, 18 Jun 2001 11:19:42 -0400 From: James D. Freels <[EMAIL PROTECTED]> To: debian-beowulf@lists.debian.org I have a set of 4 debian machines: 2 are intel, 2 are alpha. One of the intels (which would be the master machine) runs the unstable (woody) version, the other three run stable (potato) version. All machines have the 6.3.2 version of lam2, lam2-dev, and lam-runtime installed. The stable machines are at -3, while the unstable are at -11. lamboot runs fine and verifies all 4 machines to be running the lamd deamon and ready to execute. I have tried two simple test problems, both of which have the same problem. The first is the standard hello.f modified to use mpi to return the hello message from the remote machines. This program (hello) has run successfully on a "production" parallel environment so I have confidence it is not the problem. The second problem is the "trivial.f" program included in the lam2-dev examples tree. Both hello and trivial have the same problem, so I seek help here in this mailing list. When compiled, I get the following warning messages: make /usr/lib/lam/bin/hf77 -O -o trivialf trivial.f /usr/lib/gcc-lib/i386-linux/2.95.4/libg2c.a(open.o): In function `f_open': open.o(.text+0x647): the use of `tempnam' is dangerous, better use `mkstemp' However, it does produce an executable. So, when executed, I get the following error message: mpirun trivialf mpirun: cannot parse: Exec format error Since the only thing these programs do is write a message to the screen, I suspect an incompatibility between the lam packages and the current gcc compiler under woody. However, I hesitate to file a bug report until I get feedback from here. Thanks for help.. -- James D. Freels, P.E._i, Ph.D. Oak Ridge National Laboratory [EMAIL PROTECTED] - work [EMAIL PROTECTED] - home ------------------------------------------------------- -- James D. Freels, P.E._i, Ph.D. Oak Ridge National Laboratory [EMAIL PROTECTED] - work [EMAIL PROTECTED] - home
