QuantLib Grid Computing

classic Classic list List threaded Threaded
1 message Options
Reply | Threaded
Open this post in threaded view
|

QuantLib Grid Computing

Eric Ehlers-2
Hi All,

A proprietary implementation of grid computing for QuantLib was
implemented and it has been decided to release this code under the
QuantLib license.

The module depends on QuantLibAddin and was last updated for
QuantLibAddin version 0.9.0 so I have uploaded the tarball
(QLGrid-0.9.0.tar.gz) to the corresponding section of the SourceForge
downloads page:

http://sourceforge.net/project/showfiles.php?group_id=12740&package_id=159049&release_id=570898

The design of the application was specific to a particular environment,
and the documentation (included in the tarball) is sparse, so someone
coming with no background on the project would need to do a lot of work
in order to get a grid running, but the code is available to anyone who
might want to try.  I would be happy to provide support as best I can
via email.

Below is a brief summary of the application.

Regards,
Eric

- The grid runs on Linux using Sun Grid Engine.

- The client is written in platform neutral C++ and builds are provided
for Linux and Windows.

- The client transmits to the grid an XML message describing a list of
PV calculations, the grid performs the calculations and sends the list
of PVs back to the client.  Communication between client and grid is
performed via SOAP under SSL encryption.

- The smallest unit of work is a PV calculation, performed on each node
of the grid.  The calculation engine, including a complete copy of
QuantLib, is installed on each node.  The grid is stateless - the node
loads a deal, calculates and returns a price, and then waits for the
next job.

- The grid is best suited for environments where multiple PV
calculations run in parallel, for example pricing portfolios and
performing sensitivity analysis.

- The grid supports all instruments and market data objects which
QuantLib supports.  Each unit of work is simply a stream of XML
containing serialized objects to be priced as explained at
http://quantlib.org/quantlibaddin/serialization.html.

===================================================
Eric Ehlers
nazcatech sprl | Brussels | http://www.nazcatech.be
* Distributed computing for pricing analytics
* Use Microsoft Excel as a client to the Grid


------------------------------------------------------------------------------
Open Source Business Conference (OSBC), March 24-25, 2009, San Francisco, CA
-OSBC tackles the biggest issue in open source: Open Sourcing the Enterprise
-Strategies to boost innovation and cut costs with open source participation
-Receive a $600 discount off the registration fee with the source code: SFAD
http://p.sf.net/sfu/XcvMzF8H
_______________________________________________
QuantLib-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/quantlib-users