|
Parallel mpirun non interactive run uses wrong lib on cluster nodes
I am having this issue with Intel compiled version of openMpi which apparently prevents parallel running on home made cluster nodes:
Here below is the output of ldd command on the program which is on shared nfs dir:
On the Master
pier@mindthegap /mnt/sda4/WRFiFort/WRFV3/test/em_real $ ldd /mnt/sda4/WRFiFort/WRFV3/test/em_real/wrf.exe
linux-vdso.so.1 => (0x00007fffa43ff000)
libcurl.so.4 => /mnt/sda4/WRFiFort/lib/libcurl.so.4 (0x00007f46f7f9e000)
libjasper.so.1 => /mnt/sda4/WRFiFort/lib/libjasper.so.1 (0x00007f46f7d14000)
libmpi_f90.so.0 => /mnt/sda4/WRFiFort/lib/libmpi_f90.so.0 (0x00007f46f7b10000)
libmpi_f77.so.0 => /mnt/sda4/WRFiFort/lib/libmpi_f77.so.0 (0x00007f46f78d4000)
libmpi.so.0 => /mnt/sda4/WRFiFort/lib/libmpi.so.0 (0x00007f46f7605000)
libopen-rte.so.0 => /mnt/sda4/WRFiFort/lib/libopen-rte.so.0 (0x00007f46f7383000)
libopen-pal.so.0 => /mnt/sda4/WRFiFort/lib/libopen-pal.so.0 (0x00007f46f7108000)
libdl.so.2 => /lib/libdl.so.2 (0x00007f46f6f04000)
libnsl.so.1 => /lib/libnsl.so.1 (0x00007f46f6cec000)
libutil.so.1 => /lib/libutil.so.1 (0x00007f46f6ae9000)
libm.so.6 => /lib/libm.so.6 (0x00007f46f6866000)
libiomp5.so => /opt/intel/Compiler/11.1/072/lib/intel64/libiomp5.so (0x00007f46f8293000)
libpthread.so.0 => /lib/libpthread.so.0 (0x00007f46f664a000)
libc.so.6 => /lib/libc.so.6 (0x00007f46f62f1000)
libgcc_s.so.1 => /lib/libgcc_s.so.1 (0x00007f46f60da000)
libidn.so.11 => /usr/lib/libidn.so.11 (0x00007f46f5ea7000)
libssh2.so.1 => /usr/lib/libssh2.so.1 (0x00007f46f5c80000)
libssl.so.0.9.8 => /usr/lib/libssl.so.0.9.8 (0x00007f46f5a2f000)
libcrypto.so.0.9.8 => /usr/lib/libcrypto.so.0.9.8 (0x00007f46f56a5000)
librt.so.1 => /lib/librt.so.1 (0x00007f46f549c000)
libimf.so => /opt/intel/Compiler/11.1/072/lib/intel64/libimf.so (0x00007f46f5108000)
libsvml.so => /opt/intel/Compiler/11.1/072/lib/intel64/libsvml.so (0x00007f46f4ef2000)
libintlc.so.5 => /opt/intel/Compiler/11.1/072/lib/intel64/libintlc.so.5 (0x00007f46f4db4000)
libjpeg.so.8 => /usr/lib/libjpeg.so.8 (0x00007f46f4b7b000)
libifport.so.5 => /opt/intel/Compiler/11.1/072/lib/intel64/libifport.so.5 (0x00007f46f4a42000)
libifcore.so.5 => /opt/intel/Compiler/11.1/072/lib/intel64/libifcore.so.5 (0x00007f46f47cb000)
/lib64/ld-linux-x86-64.so.2 (0x00007f46f8247000)
libz.so.1 => /lib/libz.so.1 (0x00007f46f45b5000)
non interactively on node1
pier@mindthegap /mnt/sda4/WRFiFort/WRFV3/test/em_real $ ssh node1 ldd /mnt/sda4/WRFiFort/WRFV3/test/em_real/wrf.exe
Warning: untrusted X11 forwarding setup failed: xauth key data not generated
Warning: No xauth data; using fake authentication data for X11 forwarding.
linux-vdso.so.1 => (0x00007fff73ae2000)
libcurl.so.4 => /usr/lib/libcurl.so.4 (0x00007f178d2f5000)
libjasper.so.1 => /usr/lib/libjasper.so.1 (0x00007f178d09b000)
libmpi_f90.so.0 => /usr/lib/libmpi_f90.so.0 (0x00007f178ce97000)
libmpi_f77.so.0 => /usr/lib/libmpi_f77.so.0 (0x00007f178cc62000)
libmpi.so.0 => /usr/lib/libmpi.so.0 (0x00007f178c9c9000)
libopen-rte.so.0 => /usr/lib/libopen-rte.so.0 (0x00007f178c76b000)
libopen-pal.so.0 => /usr/lib/libopen-pal.so.0 (0x00007f178c50a000)
libdl.so.2 => /lib/libdl.so.2 (0x00007f178c306000)
libnsl.so.1 => /lib/libnsl.so.1 (0x00007f178c0ee000)
libutil.so.1 => /lib/libutil.so.1 (0x00007f178beeb000)
libm.so.6 => /lib/libm.so.6 (0x00007f178bc68000)
libiomp5.so => /lib64/libiomp5.so (0x00007f178d589000)
libpthread.so.0 => /lib/libpthread.so.0 (0x00007f178ba4c000)
libc.so.6 => /lib/libc.so.6 (0x00007f178b6f3000)
libgcc_s.so.1 => /lib/libgcc_s.so.1 (0x00007f178b4dc000)
librt.so.1 => /lib/librt.so.1 (0x00007f178b2d3000)
libssl.so.0.9.8 => /usr/lib/libssl.so.0.9.8 (0x00007f178b082000)
libcrypto.so.0.9.8 => /usr/lib/libcrypto.so.0.9.8 (0x00007f178acf8000)
libz.so.1 => /lib/libz.so.1 (0x00007f178aae2000)
libjpeg.so.8 => /usr/lib/libjpeg.so.8 (0x00007f178a8a9000)
libgfortran.so.3 => /usr/lib/gcc/x86_64-pc-linux-gnu/4.3.4/libgfortran.so.3 (0x00007f178a5ce000)
/lib64/ld-linux-x86-64.so.2 (0x00007f178d53d000)
interactively on node1:
pier@node1 /mnt/sda4/WRFiFort/WRFV3/test/em_real $ ldd /mnt/sda4/WRFiFort/WRFV3/test/em_real/wrf.exe
linux-vdso.so.1 => (0x00007fffafcce000)
libcurl.so.4 => /mnt/sda4/WRFiFort/lib/libcurl.so.4 (0x00007f46d7f79000)
libjasper.so.1 => /mnt/sda4/WRFiFort/lib/libjasper.so.1 (0x00007f46d7cef000)
libmpi_f90.so.0 => /mnt/sda4/WRFiFort/lib/libmpi_f90.so.0 (0x00007f46d7aeb000)
libmpi_f77.so.0 => /mnt/sda4/WRFiFort/lib/libmpi_f77.so.0 (0x00007f46d78af000)
libmpi.so.0 => /mnt/sda4/WRFiFort/lib/libmpi.so.0 (0x00007f46d75e0000)
libopen-rte.so.0 => /mnt/sda4/WRFiFort/lib/libopen-rte.so.0 (0x00007f46d735e000)
libopen-pal.so.0 => /mnt/sda4/WRFiFort/lib/libopen-pal.so.0 (0x00007f46d70e3000)
libdl.so.2 => /lib/libdl.so.2 (0x00007f46d6edf000)
libnsl.so.1 => /lib/libnsl.so.1 (0x00007f46d6cc7000)
libutil.so.1 => /lib/libutil.so.1 (0x00007f46d6ac4000)
libm.so.6 => /lib/libm.so.6 (0x00007f46d6841000)
libiomp5.so => /opt/intel/Compiler/11.1/072/lib/intel64/libiomp5.so (0x00007f46d826e000)
libpthread.so.0 => /lib/libpthread.so.0 (0x00007f46d6625000)
libc.so.6 => /lib/libc.so.6 (0x00007f46d62cc000)
libgcc_s.so.1 => /lib/libgcc_s.so.1 (0x00007f46d60b5000)
libidn.so.11 => /usr/lib/libidn.so.11 (0x00007f46d5e82000)
libssh2.so.1 => /usr/lib/libssh2.so.1 (0x00007f46d5c5b000)
libssl.so.0.9.8 => /usr/lib/libssl.so.0.9.8 (0x00007f46d5a0a000)
libcrypto.so.0.9.8 => /usr/lib/libcrypto.so.0.9.8 (0x00007f46d5680000)
librt.so.1 => /lib/librt.so.1 (0x00007f46d5477000)
libimf.so => /opt/intel/Compiler/11.1/072/lib/intel64/libimf.so (0x00007f46d50e3000)
libsvml.so => /opt/intel/Compiler/11.1/072/lib/intel64/libsvml.so (0x00007f46d4ecd000)
libintlc.so.5 => /opt/intel/Compiler/11.1/072/lib/intel64/libintlc.so.5 (0x00007f46d4d8f000)
libjpeg.so.8 => /usr/lib/libjpeg.so.8 (0x00007f46d4b56000)
libifport.so.5 => /opt/intel/Compiler/11.1/072/lib/intel64/libifport.so.5 (0x00007f46d4a1d000)
libifcore.so.5 => /opt/intel/Compiler/11.1/072/lib/intel64/libifcore.so.5 (0x00007f46d47a6000)
/lib64/ld-linux-x86-64.so.2 (0x00007f46d8222000)
libz.so.1 => /lib/libz.so.1 (0x00007f46d4590000)
How to make the program use the "right" intel libs (preventing mixing of mpi intel-gnu compiled libs) when run non interactively by mpirun?
Thanks for any hint on this,
Pier
Last edited by Pier; 05-21-2010 at 04:49 AM.
|