temcom/math-roll

SDSC math roll

SDSC "math" roll

Overview

This roll bundles a collection of Math packages: eigen, GSL, LAPACK, Parmetis,petsc, ScaLAPACK, slepc, SPRNG, sundials, and superLU.

For more information about the various packages included in the math roll please visit their official web pages:

• eigen is a C++ template library for linear algebra: matrices, vectors, numerical solvers, and related algorithms.
• GSL is a numerical library for C and C++ programmers.
• LAPACK provides routines for solving systems of simultaneous linear equations, least-squares solutions of linear systems of equations, eigenvalue problems, and singular value problems.
• ParMETIS is an MPI-based parallel library that implements a variety of algorithms for partitioning unstructured graphs, meshes, and for computing fill-reducing orderings of sparse matrices.
• petsc is a suite of data structures and routines for the scalable (parallel) solution of scientific applications modeled by partial differential equations
• ScaLAPACK is a library of high-performance linear algebra routines for parallel distributed memory machines.
• slepc is a software library for the solution of large scale sparse eigenvalue problems on parallel computers
• sundials solves nonlinear and diffential equations
• SPRNG is a scalable package for parallel pseudo random number generation which will be easy to use on a variety of architectures, especially in large-scale parallel Monte Carlo applications.
• SuperLU is a general purpose library for the direct solution of large, sparse, nonsymmetric systems of linear equations on high performance machines.

Requirements

To build/install this roll you must have root access to a Rocks development machine (e.g., a frontend or development appliance).

If your Rocks development machine does not have Internet access you must download the appropriate math source file(s) using a machine that does have Internet access and copy them into the src/<package> directories on your Rocks development machine.

Dependencies

The sdsc-roll must be installed on the build machine, since the build process depends on make include files provided by that roll.

The roll sources assume that modulefiles provided by SDSC compiler, python, and cmake rolls are available, but it will build without them as long as the environment variables they provide are otherwise defined.

The build process requires the MKL and GMP libraries and assumes that the modulefiles provided by the SDSC compiler rolls are available. It will build without the modulefiles as long as the environment variables they provide are otherwise defined.

Building

To build the math-roll, execute this on a Rocks development machine (e.g., a frontend or development appliance):

% make 2>&1 | tee build.log

A successful build will create the files math-*.iso. If you built the roll on a Rocks frontend, proceed to the installation step. If you built the roll on a Rocks development appliance, you need to copy the roll to your Rocks frontend before continuing with installation.

This roll source supports building with different compilers and for different MPI flavors. The ROLLCOMPILER and ROLLMPI make variables can be used to specify the names of compiler and MPI modulefiles to use for building the software, e.g.,

make ROLLCOMPILER=intel ROLLMPI=mvapich2_ib 2>&1 | tee build.log

The build process recognizes "gnu", "intel" or "pgi" as the value for the ROLLCOMPILER variable; any MPI modulefile name may be used as the value of the ROLLMPI variable. The default values are "gnu" and "rocks-openmpi".

For gnu compilers, the roll also supports a ROLLOPTS make variable value of 'avx', indicating that the target architecture supports AVX instructions.

Installation

To install, execute these instructions on a Rocks frontend:

% rocks add roll *.iso
% rocks enable roll math
% cd /export/rocks/install
% rocks create distro
% rocks run roll math | bash

In addition to the software itself, the roll installs individual environment module files for each tool in:

/opt/modulefiles/applications

Testing

The math-roll includes a test script which can be run to verify proper installation of the roll documentation, binaries and module files. To run the test scripts execute the following command(s):

% /root/rolltests/math.t