By Vivien Marx

GPU provider Nvidia has launched a web portal that it hopes will serve as a home base for researchers looking to access bioinformatics applications that have been optimized for graphics processing units.

The Tesla Bio Workbench, named after one of the company's GPU processors, houses downloadable GPU-optimized code for seven life science applications — AMBER, GROMACS, HOOMD, LAMMPS, NAMD, TeraChem, and VMD. The company said that by Feb. 1 it plans to add code for GROMOS, GPU-HMMER, MUMmerGPU, and CUDA-SmithWaterman.

The site also includes release notes, installation guidelines, benchmark data, publications, and — key to the effort for Nvidia — information about applicable Nvidia hardware configurations for each software package. This information is integrated with the company's e-commerce capabilities so that potential customers can purchase GPU-enabled workstations or clusters online.

The portal stems from a "critical mass" of GPU-enabled applications in the life sciences, Sumit Gupta, senior product manager for Nvidia's Tesla product group, told BioInform.

The site is part of the firm's strategy to be more visible among early adopters of GPU technology and to support programming efforts relating to GPUs.

Although traditionally at home in the gaming industry, Nvidia has found that the life science market is a breeding ground for early adopters of GPU technology. Gupta said that -omics researchers approached the company about porting code to GPUs before the firm even had a business development manager reaching out to the life sciences.

Life scientists get a "gleam" in their eyes when they are presented with the chance to power their experiments without waiting for supercomputing time, Gupta said. "When you see that, you know there is an opportunity."

Nvidia claims that its Tesla processors offer equivalent performance to quad-core CPUs with about 5 percent of the power requirements and 10 percent of the cost. Later this year, the company will release Fermi, a new processor that is expected to be twice as fast as the Tesla.

Nvidia believes that investing in the portal and pushing into the life sciences stands to grow the GPU user community and enhance the company's visibility. "It’s a community hungry for the capabilities that GPUs can provide them," Gupta said.

Gupta added that the life science market is a potentially large opportunity for the company. NAMD users alone total around 60,000 users, he said, and the total number of users for all the applications in the portal likely numbers in the hundreds of thousands.

"Eventually, when the dust settles, every biologist will have a desktop with GPUs in it and every research lab will have a GPU-enabled cluster where they can run these larger simulations," he said.

GPUs with Legs

With the launch of the portal, Nvidia is "trying to demonstrate that GPU computing, specifically Nvidia GPU computing, has legs," Joe Landman, founder and CEO of Scalable Informatics, told BioInform via e-mail. "That is, that there is an ecosystem developing around their platform."

Indeed, Gupta noted that the portal was developed to show that GPU-enabled software is plentiful and not just one or two packages by "some wacko guys in the corner."

The biology community is not yet "aware of what GPUs can enable for them," he said. "The thing we are doing new and different here is that we are now front and center promoting GPUs in this community," he said.