The Stony Brook HIV PR research project team saved time by running simulations on Itanium-based servers. The simulations took about 20,000 SGI Altix CPU hours and the project lasted about six months. The team estimated that it would have taken about ten years to complete the work using Stony Brook’s 66-node Beowulf cluster.
Challenge
• Close the gap between experimental and simulated results. Increase understanding and confidence through cross-correlation of results.
• Simulate the interaction between HIV strains and drug treatments. Find an alternative that can complete the work faster than the ten year estimated time using Stony Brook’s 66-node cluster.
Solution
• Optimize the simulation. Introduce new modeling methods to increase accuracy and parallelize the code to make it run faster.
• Get support from the National Center for Supercomputing Applications (NCSA). Run simulations on NCSA’s SGI* Altix* server cluster based on the Intel® Itanium® 2 processor.
Click here to read the full paper.