This week, Golden Helix announced that the next version of its SNP & Variation Suite software will be able to run on off-the-shelf graphics processing units and detect copy number variants in less time and at a fraction of the cost of compute clusters and cloud-based computing.
The new GPU capabilities will be included in version 7.4 of SVS that will be released in November.
For the first version of the GPU-enabled software, the company has sped up SVS's copy number variation detection algorithm, which now runs "10 to 20 times faster on a GPU than on a similarly priced central processing unit" Gabe Rudy, vice president for product development at Golden Helix, said in a statement.
GPUs can contain hundreds of microprocessors and are inexpensive when compared to clusters and clouds, Golden Helix said, adding that "extremely fast GPUs cost less than $500" and are currently being used in other markets such as oil exploration and graphic design.
Christophe Lambert, Golden Helix's president and CEO, said that "exploiting the power of GPUs" is a way to "keep bioinformatics on the desktop or small server" so that genetic analysis research remains "feasible for … researchers who are fiscally constrained" and don’t have the "resources for massive clusters and petabytes of disk storage."