Difference: DocuMentacoes (1 vs. 14)

1. WebHome. 2. DocuMentacoes. 3. HistoricoLabiocomp

MPI Alternatives
----------------
This package uses the Debian Alternatives system to allow you to choose
between the different MPI implementations.

You can check which version of MPI you use with
update-alternatives --list mpirun
update-alternatives --list mpi (for the development files)

Use 'update-alternatives --display' to list all installed implementations,
and 'update-alternatives --config' to configure which implementation to use.

update - alternatives --config mpirun para fazer o setup do mpirun (usar este).

__________________________________________________________________________________________________________________________________________________

-- RodrigoZucoloto - 02 Nov 2011

5) MPICH2 - Debian e RedHat?

Copiado de http://www.flaviotorres.com.br/fnt/artigos/mpich2.php

Por: Flavio Torres - ftorres[@]ymail.com
Publicado em: 03/08/2007

Instalando MPICH2 em Sistemas Linux.

MPICH é uma das implementações existentes para o padrão MPI (Message-Passing Interface) de bibliotecas de passagem de mensagem.

Além da biblioteca MPI, MPICH contém um ambiente de programaçao que inclui um conjunto de bibliotecas para análise de performance (profiling) de programas MPI e uma interface gráfica para todas as ferramentas.

Em outras palavras, com o MPI é possível você ter um único processo sendo executado em múltiplos servidores, um Cluster.

Este é um segundo artigo, para aqueles que não possuem um ubuntu dapper, ou brigaram muito com o python 2.3

Arquivos necessários:

* gcc
* cpp
* libc6
* lib6c-dev
* g77
* g++
* Python 2.2 ou superior

Instalando os arquivos necessários:

apt-get install gcc cpp libc6 libc6-dev g77 g++

Python, em 99% das instalações já vem na versão 2.4 ou 2.5.

Obtendo o tarball do mpich2, site do projeto: http://www-unix.mcs.anl.gov/mpi/mpich/

wget http://www-unix.mcs.anl.gov/mpi/mpich/downloads/mpich2-1.0.5p4.tar.gz

Descompacte o arquivo dentro de seu home:

tar -xvzf mpich2-1.0.5p4.tar.gz ; cd mpich2-1.0.5p4

Compile e instale:

./configure -prefix=/home/you/mpich2-install |& tee configure.log
make |& tee make.log
make install |& tee install.log

Se não utilizar um prefix, o default será /usr/local/bin.

Adicione o local de instalação em seu $PATH;

Para csh e tcsh:

setenv PATH /home/you/mpich2-install/bin:$PATH

Para Bash e sh:

export PATH=/home/you/mpich2-install/bin:$PATH

Checando se está tudo em ordem:

which mpd
which mpiexec
which mpirun

O which deve te retornar o local de instalação dos executáveis.

Após instalar em todos os hosts, devemos configurar os nomes para a resolução, edite o seu /etc/hosts e configure todas as máquinas:

vi /etc/hosts
192.168.0.1 host1
192.168.0.5 host2
192.168.0.2 host3

Agora, devemos configurar o ssh para conexão sem senha, entre todos os hosts.

Gerando a chave, lembre-se de fazer em todos os servidores:

ssh-keygen -t dsa -b 1024
* Não digite a senha quando for questionado, tecle <enter>

Agora devemos configurar o ssh para autenticar sem senha:

Passo1) Copiando a chave do host1 para o host2 e host3:

host1$ scp .ssh/id_dsa.pub usuario@host2:
host1$ scp .ssh/id_dsa.pub usuario@host3:

Passo2) Configurando a chave para autenticação no host2:

host2$ cat id_dsa.pub >> .ssh/authorized_keys
host2$ chown 600 .ssh/authorized_keys

Passo3) Configurando a chave para autenticação no host3:

host3$ cat id_dsa.pub >> .ssh/authorized_keys
host3$ chown 600 .ssh/authorized_keys

Agora, repita os 3 Passos para as 3 máquinas, no final de tudo, você deve realizar ssh sem senha da máquina:

host1 > host2 e host3
host2 > host1 e host3
host3 > host2 e host1

Configurando os arquivos do mpi que são:

* mpd.conf
* mpd.hosts

O arquivo mpd.conf contém a informação de autenticação do mpi entre as máquinas, por isto,a senha deve ser a MESMA em todos os hosts.

Se você está utilizando o root para realizar os testes, este arquivo deve estar dentro de /etc, caso esteja utilizando algum usuário comum, este arquivo deve estar dentro de $HOME.

Adicionando a senha do mpd.conf e copiando para as outras máquinas:

host1$ echo "MPD_SECRETWORD=mr45-j9z" > .mpd.conf
host1$ chmod 600 .mpd.conf
host1$ scp .mpd.conf host2:
host1$ scp .mpd.conf host3:

O arquivo mpd.hosts contém as máquinas que fazem parte do cluster para ESTE usuário;

Adicionando as máquinas do cluster no arquivo .mpd.hosts, e replicando para as outras máquinas:

host1$ echo "host1" > .mpd.hosts
host1$ echo "host2" >> .mpd.hosts
host1$ echo "host3" >> .mpd.hosts
host1$ scp .mpd.hosts host2:
host1$ scp .mpd.hosts host3:

Pronto, a parte chata já passou

Agora vamos iniciar o daemon do mpi, com o mpdboot:

host1$ mpdboot -n 3 -f .mpd.hosts
host1$ mpdtrace
host1
host2
host3

Perfeito, estão todos respondendo!!

Agora é só brincar com um teste simples:

host1$ mpiexec -n 5 mpich2-1.0.5p4/examples/cpi
Process 0 of 5 is on host1
Process 2 of 5 is on host2
Process 1 of 5 is on host3
Process 4 of 5 is on host3
Process 3 of 5 is on host1
pi is approximately 3.1415926544231230, Error is 0.0000000008333298

wall clock time = 0.925560

3) Arquivo de configuração do DHCP

/etc/dhcp/dhcpd.conf (Aqui estão as configurações que ligam os IPs com seus respstivos MACs)

/etc/init.d/isc-dhcp-server restart (Comando para reiniciar o serviço sem a necessidade de reiniciar o sistema todo)

__________________________________________________________________________________________________________________________________________________

4) Clustal-mpi

4a) README CLUSTAL-MPI

***************************

************************************************

CLUSTALW-MPI: ClustalW? Analysis Using Grid and Parallel Computing

based on ClustalW? , the multiple sequence alignment program
(version 1.82, Feb 2001)

*****************************************************************************

This README contains the help with INSTALLATION

ClustalW? is a popular tool for multiple sequence alignment. The
alignment is achieved via three steps: pairwise alignment,
guide-tree generation and progressive alignment. ClustalW? -MPI is an
MPI implementation of ClustalW? . Based on
version 1.82 of the original ClustalW? , both the pairwise
and progressive alignments are parallelized with MPI, a
popular message passing programming standard.

ClustalW? -MPI is freely available to the user community.

The software is available at
http://www.bii.a-star.edu.sg/software/clustalw-mpi/

The original ClustalW? /ClustalX can be found at ftp://ftp-igbmc.u-strasbg.fr.

Please send bug reports, comments etc. to "kuobin@bii.a-star.edu.sg".

INSTALLATION (for Unix/Linux)
------------

This is an extremely quick installation guide.

1. Make sure you have MPICH or LAM installed on your system.

2. Unpack the package in any working directory:

tar xvfp clustalw-mpi-0.1.tar.gz

3. Take a look at the Makefile and make the modifications that you might desire,
in particular:

CC = mpicc
CFLAGS = -c -g

or

CFLAGS = -c -O3

4. Build the whole thing simply by typing "make".

5. If you wanted to use serial codes to compute the neighbor-joining tree,
you would have to define the macro "SERIAL_NJTREE" when compiling trees.c:

CFLAGS = -c -g -DSERIAL_NJTREE

This macro is defined in the default Makefile. That is, to use
MPI codes in neighbor-joining tree, you have to "undefine" the
macro "SERIAL_NJTREE" in your Makefile.


SAMPLE USAGE (for Unix/Linux)
------------

1. To make a full multiple sequence alignment:
(using one master node and 4 computing nodes)

%mpirun -np 5 ./clustalw-mpi -infile=dele.input
%mpirun -np 5 ./clustalw-mpi -infile=CFTR.input

2. To make a guide tree only:

%mpirun -np 5 ./clustalw-mpi -infile=dele.input -newtree=dele.mytree
%mpirun -np 5 ./clustalw-mpi -infile=CFTR.input -newtree=CFTR.mytree

3. To make a multiple sequence alignment out of an existing
tree:

%mpirun -np 5 ./clustalw-mpi -infile=dele.input -usetree=dele.mytree
%mpirun -np 5 ./clustalw-mpi -infile=CFTR.input -usetree=CFTR.mytree

4. The environment variable, CLUSTALG_PARALLEL_PDIFF, could be used to
run the progressive alignment based on the parallelized pdiff().

By default the variable CLUSTALG_PARALLEL_PDIFF is not set, and
the progressive alignment will be parallelized accroding the structure
of the neighbor-joining tree. However, parallelized pdiff() will still
be used in the later stage when prfalign() tries to align more distant
sequences to the profiles. If you don't understand this,
simply leave the variable unset.

KNOWN PROBLEM
------------
1. On Intel IA32 platforms, slightly different neighbor-joining trees
might be obtained with and without enabling the compiler's optimization flags.

This is due to the fact that Intel processors use 80-bit FPU registers
to cache "double" variables, which are supposed to be 64-bit long. With '-O1'
or above optimizer flag, the compiler would not always immediately save the
variables involved in a double operation back to memory. Instead, intermediate
results will be saved in registers, having 80-bit of precision. This would
cause problem for nj_tree() because it is sensitive to the precision of floating
point numbers.

Solutions:

(1) Other platforms, including Intel's IA64, don't seem to have this problem.

or

(2) Building "trees.c" with optios like the below: (potentially
with high performance overhead)

%gcc -c -O3 -ffloat-store trees.c // GNU gcc

%icc -c -O3 -mp trees.c // Intel C compiler

or

(3) Decalring relevant variables as "volatile" in nj_tree():

volatile double diq, djq, dij, d2r, dr, dio, djo, da;
volatile double *rdiq;

rdiq = (volatile double *)malloc(((last_seq-first_seq+1)+1)*
sizeof(volatile double));
...
...
free((void)rdiq);

4b) Manual de usdo extraído de: http://www.cuhk.edu.hk/itsc/compenv/research-computing/organon/cwmpi.html#ii

OBS: o clustalw-mpi já está instalado em todo o cluster.


1. Introduction of ClustalW? -MPI

ClustalW? is a general-purpose multiple sequence alignment program for DNA or proteins.

The alignment is achieved via three steps:

• pairwise alignment;
• guide-tree generation; and
• progressive alignment.

Job Monitoring

List the current jobs on the cluster
%qstat

List the currently running jobs on the cluster
%qstat –a

Lists nodes allocated to running jobs
%qstat –r

Show detailed information on a specific job
% qstat –n

Show detailed information for all queues
%qstat -f <jobid>

Advance PBS command:

Advanced users can type % man q_command to see the details of the following command.

Guilherme:

Proxy:
add no /etc/profile das estações (internet)
export http_proxy=http://192.168.1.1:3128

NFS
add no /etc/fstab (monta home NFS)
darthvader.bio.intranet.ufba.br:/home /home nfs rw,hard,intr 0 0

Torque nas estações:
/var/spool/torque/server_name > uma única linha com o nome do servidor "darthvader"
/var/spool/torque/mom_priv/config > adicionar nome do servidor "darthvader"

# apt-get install libnss-ldap libpam-ldap nscd

LDAP server URI: ldap://192.168.1.2/
Distinguished name of the search base: dc=republica,dc=star,dc=wars
LDAP version to use: 3
LDAP account for root: cn=admin,dc=republica,dc=star,dc=wars

-- RodrigoZucoloto - 02 Nov 2011