On Tue, 2009-04-21 at 08:20 -0400, Adam C Powell IV wrote: > On Sun, 2009-04-19 at 18:21 -0430, Muammar El Khatib wrote: > > Hi Adam, > > > > I don't mind if you do a NMU for this. It is a little bit difficult for me > > at > > this time to make the changes because I am writing my thesis, and finishing > > NM > > process (which is like another career :)) > > > > So, go ahead with this NMU. I do apologize. > > Thanks very much. To be clear, I plan to replace the current mpich and > lam packages with just one set build-depending on mpi-default-dev, with > the -dev binary package depending on that as well, and with the -test > package recommending mpi-default-bin. > > I'll probably do this on Friday, so let me know before then if you have > a problem with this plan.
I'm behind schedule, just finished this for blacs-mpi, result is in http://lyre.mit.edu/~powell/mumps/ . All tests pass except for c-shared: % mpirun -np 1 /usr/lib/blacs/cblacs_test_shared-openmpi BLACS WARNING 'No need to set message ID range due to MPI communicator.' from {-1,-1}, pnum=0, Contxt=-1, on line 18 of file 'blacs_set_.c'. [workhorse:31996] *** Process received signal *** [workhorse:31996] Signal: Segmentation fault (11) [workhorse:31996] Signal code: Address not mapped (1) [workhorse:31996] Failing at address: 0x3ba25cf8 [workhorse:31996] [ 0] /lib/libc.so.6 [0x7fb53aac4f60] [workhorse:31996] [ 1] /usr/lib/libmpi.so.0(PMPI_Comm_group+0x50) [0x7fb53b7e4a60] [workhorse:31996] [ 2] /usr/lib/libblacs-openmpi.so.1(BI_TransUserComm+0x25) [0x7fb53ba4ca45] [workhorse:31996] [ 3] /usr/lib/libblacs-openmpi.so.1(Cblacs_gridmap+0x132) [0x7fb53ba3e5f2] [workhorse:31996] [ 4] /usr/lib/libblacs-openmpi.so.1(Cblacs_gridinit+0x2ea) [0x7fb53ba421da] [workhorse:31996] [ 5] /usr/lib/blacs/cblacs_test_shared-openmpi [0x4036c4] [workhorse:31996] [ 6] /usr/lib/blacs/cblacs_test_shared-openmpi [0x478dcc] [workhorse:31996] [ 7] /lib/libc.so.6(__libc_start_main+0xe6) [0x7fb53aab11a6] [workhorse:31996] [ 8] /usr/lib/blacs/cblacs_test_shared-openmpi [0x403559] [workhorse:31996] *** End of error message *** mpirun noticed that job rank 0 with PID 31996 on node workhorse exited on signal 11 (Segmentation fault). (With more processes, it gives more of this same error.) Static and fortran test binaries run fine and pass all tests. Any ideas? > If you like, I can also change the maintainer to pkg-scicomp with you as > the first uploader (and me as second), but I will not do this unless I > hear from you. Did not do this, you're still maintainer and this will be an NMU. -Adam -- GPG fingerprint: D54D 1AEE B11C CE9B A02B C5DD 526F 01E8 564E E4B6 Engineering consulting with open source tools http://www.opennovation.com/
signature.asc
Description: This is a digitally signed message part
