Accelerating FaST-LMM for epistasis tests
Algorithms and Architectures for Parallel Processing: 17th International …, 2017•Springer
We introduce an enhanced version of FaST-LMM that maintains the sensitivity of this
software when applied to identify epistasis interactions while delivering an acceleration
factor that is close to 7.5 * on a server equipped with a state-of-the-art graphics coprocessor.
This performance boost is obtained from the combined effects of integrating a dictionary for
faster storage of the test results; a re-organization of the original FaST-LMM Python code;
and off-loading of compute-intensive parts to the graphics accelerator.
software when applied to identify epistasis interactions while delivering an acceleration
factor that is close to 7.5 * on a server equipped with a state-of-the-art graphics coprocessor.
This performance boost is obtained from the combined effects of integrating a dictionary for
faster storage of the test results; a re-organization of the original FaST-LMM Python code;
and off-loading of compute-intensive parts to the graphics accelerator.
Abstract
We introduce an enhanced version of FaST-LMM that maintains the sensitivity of this software when applied to identify epistasis interactions while delivering an acceleration factor that is close to 7.5 on a server equipped with a state-of-the-art graphics coprocessor. This performance boost is obtained from the combined effects of integrating a dictionary for faster storage of the test results; a re-organization of the original FaST-LMM Python code; and off-loading of compute-intensive parts to the graphics accelerator.
Springer
Showing the best result for this search. See all results