Max Planck Institute of Biochemistry uses MathWorks parallel
computing tools for cancer research
24 November 2008
The MathWorks has announced that the Max Planck Institute of
Biochemistry is using its MATLAB, Parallel Computing Toolbox, MATLAB
Distributed Computing Server, and other MathWorks tools to accelerate
its workflow and reduce research and development time.
Researchers at the department of Molecular Structural Biology at the
Max Planck Institute of Biochemistry, based in Martinsried, near Munich,
Germany, study the relationship between the structure and activity of
macromolecular protein complexes involved protein degradation in cells.
This complex research requires high-throughput tools and procedures
capable of efficiently processing vast amounts of data.
Max Planck Institute researchers turned to MathWorks tools to develop
streamlined procedures for their data-intensive applications, including
image acquisition, filtering, processing, and 3-D reconstruction. For
techniques such as pattern matching and single-particle reconstruction
that required computationally intensive algorithms, researchers used
Parallel Computing Toolbox and MATLAB Distributed Computing Server to
accelerate computation over a 64-node cluster.
“Reconstructing a 3-D volume typically takes days on a single CPU,
but by using Parallel Computing Toolbox and MATLAB Distributed Computing
Server from The MathWorks we were able to speed up our processing by 20
to 30 times,” said Andreas Korinek, scientist at the Max Planck
Institute of Biochemistry.
“Particularly helpful was the ability to use our cluster productively
from the MATLAB environment without having to be experts in parallel
programming or having to learn another programming language. The changes
to the existing serial applications are minimal; in most cases, our
researchers did not have to go beyond changing for-loops to parallel
for-loops to parallelize our MATLAB code and use the cluster
productively.”
“Intensive research projects that require complex data processing can
benefit greatly from the efficiencies that parallel computing affords.
However, parallel programming is hard and traditionally engineers and
scientists have had to program their applications using low-level
languages,” said Silvina Grad-Freilich, manager of parallel computing
and application deployment marketing at The MathWorks.
“With tools such as Parallel Computing Toolbox and MATLAB Distributed
Computing Server, The MathWorks is committed to helping MATLAB users
seamlessly make the transition to parallel programming.”
Bookmark this page