Stone Ridge Technology

Website Links

Description

As experts in science and computing, Stone Ridge Technology can identify bottleneck code in your software algorithms and provide solutions that accelerate your critical applications by as much as 100X. Visit our Toolshed for additional bioinformatics solutions resources. Stone Ridge Technology has worked with Bioinformatics algorithms for over three years on CPU, GPU and FPGA platforms. As the data produced by sequencing hardware has experienced a "Moore's Law" like growth for over a decade, the processing step has become increasingly challenging. The most popular approach to sequencing genetic data results in billions of short 25-45 base pair reads which need to be aligned to a reference genome. There are three broad classes of algorithms that accomplish this: Dynamic Programming (e.g. Smith-Waterman), hash-tables (e.g. BLAST, ZOOM) and suffix trees (e.g. Bowtie). Bioinformatics algorithms are a particularly good fit for reconfigurable computing on FPGA's because the data type is best represented in a non-standard 2 bit form and the operations are typically comparative in nature (i.e. no floating point required). Many of them can be cast into a streaming format which takes advantage of the FPGA's spatial and temporal parallelism. Stone Ridge Technology has ported both the Smith Waterman and the BLAST algorithm to FPGA hardware. The details of these efforts can be found in papers in the HPC Toolshed. Our most recent work in this area was supported by an NSF grant which allowed us to study efficient FPGA implementations of Bioinformtics algorithms. Our principle finding in this work was that FPGAs are 5x less expensive in capital costs, 20x less expensive in performance per watt and 15x denser in footprint. The work accomplished for NSF was completed using our RDX-11 hardware.

Fact sheet

Products & services

Similar companies nearby