Cluster/Tools2012. 1. 26. 17:21

installing BLAS

# tar zxvfp blas.tgz 

# cd BLAS

# make all


installing Lapack

# tar zxvfp lapack-3.4.0.tgz 

# cd lapack-3.4.0

# cp make.inc.example make.inc

# cmake ../lapack-3.4.0

# make

# make install


installing fftw 3.3

# tar zxvfp fftw-3.3.tar.gz

# cd fftw-3.3

# ./configure

# make

# make install


installing fftw 2.1.5

# wget http://www.fftw.org/fftw-2.1.5.tar.gz

# tar zxvfp fftw-2.1.5.tar.gz

# cd fftw-2.1.5

# ./configure

# make

# make install


installing ATLAS

# tar xvfp atlas3.8.3.tar 

# cd ATLAS/

# mkdir linux_x86_64

# cd linux_x86_64

# ../configure -b 64 -D c -DPentiumCPS=2216 -Fa alg -fPIC --with-netlib-lapack=$HOME/src/lapack-3.2.1/lapack_LINUX.a -Si cputhrchk 0 && make && make install


installing CPMD

#./mkconfig.sh LINUX-x86_64-FEDORA-MPI > Makefile

# vi Makefile

===================================================

LFLAGS = -L/usr/local/lib/ -llapack -lblas -lfftw

===================================================

# make


Posted by 변익수
Cluster/SGE2011. 12. 28. 09:55

아래와 같은 에러 발생 시
====================================================================================================================
libibverbs: Warning: RLIMIT_MEMLOCK is 32768 bytes.
    This will severely limit memory registrations.
libibverbs: Warning: RLIMIT_MEMLOCK is 32768 bytes.
    This will severely limit memory registrations.
libibverbs: Warning: RLIMIT_MEMLOCK is 32768 bytes.
    This will severely limit memory registrations.
--------------------------------------------------------------------------
The OpenFabrics (openib) BTL failed to initialize while trying to
allocate some locked memory.  This typically can indicate that the
memlock limits are set too low.  For most HPC installations, the
memlock limits should be set to "unlimited".  The failure occured
here:

  Local host:    compute-0-10.local
  OMPI source:   btl_openib_component.c:1066
  Function:      ompi_free_list_init_ex_new()
  Device:        mlx4_0
  Memlock limit: 32768

You may need to consult with your system administrator to get this
problem fixed.  This FAQ entry on the Open MPI web site may also be
helpful:

    http://www.open-mpi.org/faq/?category=openfabrics#ib-locked-pages
--------------------------------------------------------------------------
--------------------------------------------------------------------------
WARNING: There was an error initializing an OpenFabrics device.

  Local host:   compute-0-10.local
  Local device: mlx4_0
--------------------------------------------------------------------------
libibverbs: Warning: RLIMIT_MEMLOCK is 32768 bytes.
    This will severely limit memory registrations.
libibverbs: Warning: RLIMIT_MEMLOCK is 32768 bytes.
    This will severely limit memory registrations.
libibverbs: Warning: RLIMIT_MEMLOCK is 32768 bytes.
    This will severely limit memory registrations.
libibverbs: Warning: RLIMIT_MEMLOCK is 32768 bytes.
    This will severely limit memory registrations.
libibverbs: Warning: RLIMIT_MEMLOCK is 32768 bytes.
    This will severely limit memory registrations.
[compute-0-10.local:02981] 7 more processes have sent help message help-mpi-btl-openib.txt / init-fail-no-mem
[compute-0-10.local:02981] Set MCA parameter "orte_base_help_aggregate" to 0 to see all help / error messages
[compute-0-10.local:02981] 7 more processes have sent help message help-mpi-btl-openib.txt / error in device init
====================================================================================================================

1. limits.conf
# vi /etc/security/limits.conf
* soft memlock unlimited
* hard memlock unlimited


2. For using mpirun in SGE session
# qconf -mconf
======================================================
execd_params                 H_MEMORYLOCKED=infinity
======================================================
root@ngscluster1.local modified "global" in configuration list

'Cluster > SGE' 카테고리의 다른 글

Execution host add, remove  (0) 2011.12.16
Posted by 변익수
Cluster/SGE2011. 12. 16. 09:21

1. Adding an execution host
Make the new host an administrative host

# qconf -ah <hostname>

As root on this new host, run the following script from $SGE_ROOT

# install_execd

 

2. Removing an execution host
# qconf -shgrp @allhosts
============================================================================
group_name @allhosts
hostlist test compute-0-0 compute-0-1 compute-0-3 compute-0-4 compute-0-5 \
         compute-0-13f96093_13440deac47__8000
============================================================================

# qconf -shgrp @allhosts > test.txt

# qconf -Mhgrp test.txt
root@sge.kobic.re.kr modified "@allhosts" in host group list

# qconf -shgrp @allhosts
============================================================================
group_name @allhosts
hostlist test compute-0-0 compute-0-1 compute-0-3 compute-0-4 \
         compute-0-13f96093_13440deac47__8000
============================================================================

### Delete the host
# qconf -de compute-0-5

# qconf -dh compute-0-5

### Finally, delete the configuration for the host
# qconf -dconf compute-0-5

'Cluster > SGE' 카테고리의 다른 글

memlock limits(Infiniband 관련 에러)  (0) 2011.12.28
Posted by 변익수