What is it about?
Graphcore IPU, a hardware accelerator originally developed for artificial intelligence operations, successfully speeds up the alignment of protein and DNA molecules, making the process up to 10 times faster than state-of-the-art methods on GPUs and CPUs.
Featured Image
Why is it important?
This approach can make the alignment of protein and DNA sequences for genome assembly more efficient, which is a fundamental problem in computational biology.
Perspectives
Over the past decade, scientists have turned to graphics processing units (GPUs), originally developed to accelerate graphics rendering in video games, – to speed up sequence alignment through parallel computation. By developing IPUs for AI applications, we wanted to find out if they could use the new accelerators to tackle this problem. IPUs attracted us because they have substantial on-device bandwidth for transferring data and can handle uneven and unpredictable workloads. X-Drop, a popular algorithm for aligning sequences, has a very irregular computation pattern. If two sequences match, the algorithm requires a lot of computation to determine the correct alignment, but if they don’t match, the algorithm just terminates. GPUs struggle with this type of irregular and dynamic computation, but the IPU excelled. From an article Patricia Waldron wrote about our work for Cornell Chronicles: https://news.cornell.edu/stories/2023/11/processor-made-ai-speeds-genome-assembly
Giulia Guidi
Cornell University
Read the Original
This page is a summary of: Space Efficient Sequence Alignment for SRAM-Based Computing: X-Drop on the Graphcore IPU, November 2023, ACM (Association for Computing Machinery),
DOI: 10.1145/3581784.3607094.
You can read the full text:
Contributors
The following have contributed to this page
