From: Seán <sea...@gm...> - 2009-05-07 14:06:12
|
Seán wrote: > > I seem to have the same problem. > I have solved my initial problem. It was caused by interference of another valgrind version already installed. I did a "complete remove" using synaptic, instead of the "remove" I had previously tried. I now have libmpiwrap.so installed in the directory /usr/local/lib/valgrind/x86-linux/libmpiwrap.so. I am trying to run an MPI HelloWorld program with valgrind. I am not getting the expected output "valgrind MPI wrappers 31901: Active for pid 31901" . Instead I get: sean@seanscomputer:~/CODE$ LD_PRELOAD=/usr/local/lib/valgrind/x86-linux/libmpiwrap.so mpirun -np 2 /usr/local/bin/valgrind ./main ==2869== Memcheck, a memory error detector. ==2869== Copyright (C) 2002-2008, and GNU GPL'd, by Julian Seward et al. ==2869== Using LibVEX rev 1884, a library for dynamic binary translation. ==2869== Copyright (C) 2004-2008, and GNU GPL'd, by OpenWorks LLP. ==2869== Using valgrind-3.4.1, a dynamic binary instrumentation framework. ==2869== Copyright (C) 2000-2008, and GNU GPL'd, by Julian Seward et al. ==2869== For more details, rerun with: -v ==2869== valgrind: seanscomputer: command not found p0_2869: p4_error: Child process exited while making connection to remote process on localhost: 0 ==2869== Conditional jump or move depends on uninitialised value(s) ...... Does anyone know what that "valgrind: hostname: command not found" error is? -- View this message in context: http://www.nabble.com/valgrind-and-MPI-wrappers-tp10324423p23427517.html Sent from the Valgrind - Users mailing list archive at Nabble.com. |