Free mpi software for linux

MPI Ruby is a Ruby binding of MPI. MPI Rubys primary goal in making this binding was to make the power of MPI available to Ruby users in a way that fits into the languages object oriented model.
In order to do this, the buffer and datatype management necessary in the C, C++, and Fortran bindings have been removed. What this means is that MPI Ruby allows you to treat objects as messages.
MPI Ruby also aims to be a complete binding to MPI in that it offers access to nearly all functionality of MPI. While there is not a one-to-one correspondence to functions and constants in the Ruby and C/C++/Fortran bindings, all of the communication and topology features are available.
There are fewer methods in the Ruby binding than there are functions in the C/C++/Fortran bindings, but this is mainly due to the fact that the programmer no longer needs to deal with buffers and datatypes.
Enhancements:
- examples/irecv.rb: Removed sleep from irecv example
- configure: Removed configure
- examples/Makefile.am: Added op example to Makefile
- examples/redsubmit.rb, examples/op.rb, examples/redhalt.rb, examples/red.rb:
New examples:
User-defined operations
Ruby Execution Daemon (red).
- src/ops.rb, src/main.c, src/mpi.c, src/mpi_comm.c, src/mpi_group.c, src/mpi_keyval.c, src/mpi_op.c, src/mpi_op_fns.c, src/mpi_request.c:
Fixed all of the rb_str_new2()s that were causing marshalling problems.
Fixed defines of singleton methods.
Fixed dims_create()
Now works with MPICH because of atexit(MPI_Finalize)
Set the MPI error handler
Fixed operators in MPI::Group (+ -> | and ^ -> &)
- examples/Makefile.am: New examples.
- docs/rd/mpi_group.rd, docs/rd/mpi_keyval.rd, docs/rd/mpi_op.rd, docs/rd/mpi_ruby.rd, docs/rd/mpi_comm.rd, docs/man/man3/Makefile.am, docs/man/man3/mpi_comm.3, docs/man/man3/MPI_Ruby.3, docs/man/man3/MPI_Status.3, docs/man/man3/MPI_Exception.3, docs/man/man3/MPI_Group.3, docs/man/man3/MPI_Keyval.3, docs/man/man3/MPI_Op.3, docs/man/man3/MPI_Request.3, docs/man/man3/MPI_Comm.3:
Doc updates to reflect fixes to singleton methods in several classes.
Typos fixed.
- docs/man/man1/mpi_ruby.1, docs/man/man1/Makefile.am:
Short doc on the interpreter itself (how to run)
- docs/man/Makefile.am: Added man1
- docs/html/index.html, docs/html/mpi_comm.html, docs/html/mpi_group.html, docs/html/mpi_keyval.html, docs/html/mpi_op.html, docs/html/mpi_ruby.html:
Doc updates to reflect fixes to singleton methods in several classes.
Typos fixed.
- configure, configure.in: Bumped to 0.3
Added man1/Makefile to output

GridMPI is a new open-source free-software implementation of the standard MPI (Message Passing Interface) library designed for the Grid. GridMPI project enables unmodified applications to run on cluster computers distributed across the Grid environment.
GridMPI team found that it is feasible to connect cluster computers and to run ordinary scientific applications in distance upto 500 miles. Simple experiment has shown that most MPI benchmarks scale fine upto 20 millisecond round-trip latency which corresponds to about 500 miles in distance, when the clusters are connected by fast 1 to 10 Gbps networks. 500 miles covers the major cities between Tokyo--Osaka in Japan.
Thus, applications which are too large to run on a local cluster should run on multiple clusters in the Grid environment with acceptable performance. However, it is only feasible when using an efficient MPI implementation [1]. Existing implementations are not efficient enough mainly because of the two reasons: their focus on security features and TCP performance problems.
GridMPI skips security layers assuming dedicated secure links. The institutes housing large clusters tend to have their own networks to connect to other institutes in most cases. GridMPI so focuses on the performance on TCP. Since existing implementations are in most cases designed for MPP machines and recently clusters with special hardware, their performance on TCP with Ethernet is not optimal.
Also TCP performance itself is not optimal for the work load of the MPI traffic. In addition, support for heterogeneous combinations of computers of the existing MPI implementations is not satisfactory. Thus, GridMPI is designed and implemented from the scratch. GridMPI is carefully coded and tested with heterogeneity in mind.
Main features:
- Full conformance to the standard: GridMPI passes 100% of the functional tests of the large test suites from ANL and Intel (MPI-1.2 level).
- Full heterogeneity support: GridMPI is fully tested with combinations of processors of 32bit/64bit and big/little-endian.
- Primary support of TCP/IP and sockets: GridMPI is written from scratch and it is new and clean. It is efficient with sockets, and thus suitable for the Grid as well as ordinary Ethernet-based clusters.
- Cooperation with Grid job submission: GridMPI can be used with Globus, Unicore, tool from NAREGI project, etc.
- Checkpointing support: GridMPI supports checkpointing on Linux/IA32 platforms to restart long-running applications from failure.
- Vendor MPI support: GridMPI supports IBM-MPI, Fujitsu-Solaris-MPI, Intel-MPI, and any MPICH-based MPI for clusters with special communication hardware.
Enhancements:
- Minor bugfixes were made.

BSPonMPI is a platform independent software library for developing parallel programs. BSPonMPI implements the BSPlib standard (with one small exception) and runs on all machines which have MPI.
This last property is the main feature of this library and with this feature it distinguishes itself from other libraries such as the Oxford BSP Toolset and PUB.
What are MPI and BSPlib?
MPI stands for Message Passing Interface. This API should make it easy to write a parallel program. However in practice it is still very complicated, because the API exists of hundreds of functions. It is still like programming in a very low level programming language, e.g. assembly.
Astonishingly there exists another API which is designed for the purpose and is very simple. It consists of only 20 primitives, which provide the same functionality and speed. BSPlib, as this other API is called, allows you to write parallel programs according to the BSP programming paradigm, see e.g. Parallel Scientific Computation: A Structured Approach using BSP and MPI by Rob H. Bisseling.
This paradigm lets you program a parallel algorithm in a very structured manner, resulting in readable and fast code. BSPlib is already implemented for several supercomputers and pc clusters, but as it is less popular than MPI, it is not implemented for all hardware platforms. As engineers and mathematicians always want the last percentage of computing power, an efficient implementation on top of MPI is imperative.
Why should I use it?
Currently there are two major BSPlib implementations: Oxford BSP Toolset and PUB. Both are implemented for specific hardware platforms (Cray T3E or SGI Origin, etc...) and they have a platform independent version on top of MPI. However the architecture of their software library is optimised for the use of hardware specific features. Building on top of MPI was never their primary objective. So if your hardware/software is not supported by one of these two libraries, then you should use BSPonMPI in combination with an MPI library.
Enhancements:
- The main data structure has been rewritten, which resulted in a big performance increase.
- Now it is ready for the real world, as it offers BSP communication at almost the same speed as the Oxford BSP Toolset.
- Sometimes, it is even faster.

MP-MPICH provides a Multi-platform MPI implementation.

MP-MICH is a Multi-platform uniform MPI implementation, based on MPICH and SCI-MPICH, resulting in a high performance, consistent MPI across both ethernet and SCI networks in a hybrid environment. There is a single, standardized source tree for all platforms. It is compliant with the MPI-1 standard.

MP-MPICH uses one common source tree for this purpose. It has a new device for Windows NT/2000/2003 to communicate via sockets, and an adapted shared memory device for SMP systems. Additionally, MP-MPICH is a superset of SCI-MPICH, that means that it includes another new device to support direct communication via a fast SCI interconnect. With SCI-MPICH, inter-process message latencies for small messages (for processes running on separate nodes) below 4 us are achieved, and the current measured maximum inter-node (PingPong-)bandwidth is at about 310 MB/s.

In fact, the development of MP-MPICH started when the first release of SCI-MPICH was finished. This release was designed for Solaris and Linux, but since it uses our SMI library which is also available for Windows NT, we thought it would be fun to port SCI-MPICH to NT. The initial port took only 4 hours and we were up and running with SCI-MPICH for NT. The additional/modified devices which are necessary to use MPICH on NT without these extremely fast SCI boards required more time to develop.

PDL::Parallel::MPI Perl module contains routines to allow PDL objects to be moved around on parallel systems using the MPI library.

SYNOPSIS

use PDL;
use PDL::Parallel::MPI;
mpirun(2);

MPI_Init();
$rank = get_rank();
$a=$rank * ones(2);
print "my rank is $rank and $a is $an";
$a->move( 1 => 0);
print "my rank is $rank and $a is $an";
MPI_Finalize();

MPI STANDARD CALLS

Most of the functions from the MPI standard may be used from this module on regular perl data. This is functionallity inherited from the Parallel::MPI module. Read the documentation for Parallel::MPI to see how to use.

One may mix mpi calls on perl built-in-datatypes and mpi calls on piddles.

use PDL;
use PDL::Parallel::MPI;
mpirun(2);

MPI_Init();
$rank = get_rank();
$pi = 3.1;
if ($rank == 0) {
MPI_Send($pi,1,MPI_DOUBLE,1,0,MPI_COMM_WORLD);
} else {
$message = zeroes(1);
$message->receive(0);
print "pi is $messagen";
}
MPI_Finalize();

mpiBLAST is an MPI based parallel implementation of NCBI BLAST. The project consists of a pair of programs that replace formatdb and blastall with versions that execute BLAST jobs in parallel on a cluster of computers with MPI installed. There are two primary advantages to using mpiBLAST versus traditional BLAST.

First, mpiBLAST splits the database across each node in the cluster. Because each nodes segment of the database is smaller it can usually reside in the buffer-cache, yielding a significant speedup due to the elimination of disk I/O. Second, it allows BLAST users to take advantage of efficient, low-cost Beowulf clusters because interprocessor communication demands are low.

mpiBLAST achieves super-linear speedup in situations where the database is too large to fit into RAM, and near linear speedup in other situations. It does not require a dedicated cluster.

mpt-status is a query tool for accessing the running configuration and status of LSI SCSI HBAs. mpt-status is a heavily modified version of the original mpt-status-1.0 tool written by Matt Braithwaite.
It allows you to monitor the health and status of your RAID setup. Currently supported and tested HBAs are the LSI 1030 SCSI RAID storage controller and LSI SAS1064 SCSI RAID storage controller.
Since the tool uses the MPI (message passing interface), chances are high that the basic information regarding RAID status will be available for all LSI-based controllers.
Enhancements:
- This is the first release with proper 64-bit support and mostly clean interface querying.
- The mpt-status(8) man page has been added.
- Preliminary work has been done on integrating S.M.A.R.T information in the report output.
- This would allow one to monitor for failing hardware and preemptively replace it without business interruption.
- This information can be queried via the --newstyle/-n parameter.
- The new style parameter also provides you with the percentage status of the synchronization after a degradation.

ParallelKnoppix is a remaster of the Knoppix live CD distibution of GNU Linux that allows setting up a cluster of machines for parallel processing using the LAM-MPI and/or MPICH implementations of MPI.

You can convert a room full of machines running Windows into a Linux cluster, and when you shut down, your Windows machines are in their original state. The computers in the cluster can be homogeneous or heterogeneous.

Getting the cluster up and running takes about 5 minutes, if the machines have PXE network cards. Clusters from 2 to 200 machines are supported.