<span class="var-sub_title">Hummingbird: Efficient Performance Prediction for Executing Genomics Applications in the Cloud</span> SC18 Proceedings

The International Conference for High Performance Computing, Networking, Storage, and Analysis

Fourth Computational Approaches for Cancer Workshop (CAFCW18)


Hummingbird: Efficient Performance Prediction for Executing Genomics Applications in the Cloud

Abstract: A major drawback of executing existing genomics pipelines on cloud computing facilities is that the onus of efficiently executing it on the best configuration lies on the user. Lack of knowledge regarding which cloud configuration is best to execute a pipeline often results in an unnecessary increase in cost due to selecting a more expensive cloud tier than needed. Resources in the cloud are expensive, so determining the best configuration before actually running the pipeline saves money and time. To this end, we introduce Hummingbird, a framework that predicts the best configuration to execute genomics pipelines on Google cloud.

Archive Materials


Back to Fourth Computational Approaches for Cancer Workshop (CAFCW18) Archive Listing

Back to Full Workshop Archive Listing