Official Site Bumtion Network

I hate cleaning, who doesn’t? You have to get your hands dirty so you can clean your dishes with liquid detergent or bend down so you can mop the floor properly, but this is only the easy part compared to the work done just to make that shampoo available to you.

Cleansing products such as shampoos and liquid detergents may seem like something you make by mixing of chemicals only, but finding ways to mix them together so they can clean better while protecting the environment involves very complex computing powers as well. Traditionally, the process of finding and testing surfactants, molecules that provide the cleaning capacity and texture of cleansing products, can be very time consuming and costly but with the help of the parallel processing power of Nvidia Telsa GPUs, the process can be much more cost-effective and accurate.

“The computer models needed to accurately simulate surfactant properties are extremely demanding in terms of computational power. We discovered that by adding just two NVIDIA Tesla C1060 GPUs, each node in our newest cluster can do 16 times more work, and thus multiplies our local compute capacity far beyond what we could previously get through the national supercomputing centers.” -- Axel Kohlmeyer of the Institute for Computational Molecular Science at Temple University.

Find out more after the jump!

Using a single GPU-optimized molecular dynamics simulation on two Telsa GPUs, the researchers at Temple University are developing a simulation system that can run as fast as 128 CPU cores of Cray XT3 supercomputer. The solution will then allow large companies such as Procter and Gamble to develop products that can help us clean better and faster.

The Temple researchers are using GPU-accelerated HOOMD (Highly Optimised Object Oriented Molecular Dynamics) simulation software, written by researchers at the Department of Energy’s Ames Laboratory to leverage the NVIDIA GPUs. In addition to deploying a small local GPU cluster, the university team will also look to scale its work using the NCSA Lincoln cluster, where the computational output has been boosted to 47 TeraFLOPS through the addition of Tesla S1070 1U GPU systems.

For more information, please visit the following websites:

- Institute for Computational Molecular Science at Temple University: www.temple.edu/cst/icms/

- The HOOMD application from Ames Lab: http://codeblue.umich.edu/hoomd-blue/

- NCSA Supercomputer: www.ncsa.uiuc.edu/UserInfo/Resources/Hardware/Intel64TeslaCluster/

- Molecular dynamics using GPUs: www.nvidia.com/object/molecular_dynamics.html