Welcome to the CBIIT HPC Wiki! This page is actively under construction and will soon serve as a central hub to provide you with the latest developments and resources related to the CBIIT HPC Program. We look forward to serving you!
The High-Performance Computing Program Development effort aims to foster the expanded use of a high-performance computing ecosystem to accelerate advances in predictive oncology research and clinical applications. Both driven and enabled by the rapid growth rates of information collected and generated about cancer, the opportunity for ever increasing computational capability grows as the data is analyzed, explored and utilized to provide critical insight into cancer. The program aims to develop the computational and data science ecosystem by addressing critical needs in compute, data transfer, data management, exploration and education in these areas required to advance the mission of the NCI.
*For support inquiries, please contact us at nci-cbiit-hpc@list.nih.gov
CLICK THE BELOW IMAGE for a larger view of CBIIT's HPC Overview
Topic: Streamlined Transfer and Sharing of Large-scale Sensitive Data to Advance Cancer Research |
Speaker: Ian Foster, Ph.D. Read Dr. Foster's professional bio.
|
Presentation synopsis: Advances in genomics and data analytics create new opportunities for cancer research and personalized medical treatment via large-scale federation of genomic, clinical, imaging and other data from many thousands of patients across institutions around the world. Despite these opportunities and promising early results, cancer research is often stymied by information technology barriers. One major barrier is a lack of tools for the reliable, secure, rapid, and easy transfer, sharing, and management of large collections of human data. In the absence of such tools, security and performance concerns often prevent sharing altogether or force researchers to resort to slow and error prone shipping of physical media. If data are received, timely analysis is further impeded by the difficulties inherent in verifying data integrity and managing who can access data and for what purpose. Dr. Foster will discuss how the mature Globus data management platform addresses these obstacles to discovery and explain how its intuitive, web-based interfaces enable use by researchers without specialized IT knowledge.
|
Yuba Bhandari has developed state of art algorithm that matches experimental molecular
dynamics results with simulated code.
Eric Stahlberg has explored the optimization of this code in terms of runtime and input
interface on intel multicore processors using OpenMP.
Given the availability of GPU accelerator on Biowulf and the national labs
supercomputer, it is worth exploring if further speedup can be achieved using
GPUs.
Objectives:
1- Implement an GPU version of the time consuming hotspots of the code.
2- Explore the implementation of a web interface to allow users outside the NCI
to use the application.
3- Explore further refactoring of the code to enhance the optimization process.
CLICK HERE for the full report.
High Performance Computing (HPC) has an invaluable impact on driving advances in cancer research. Projects conducted within the context of the Precision Medicine Initiative, the National Cancer Institute, the Department of Energy pilots, and the National Strategic Computing Initiative are only examples of how new frontiers can be pushed using the next generation of supercomputers.
Areas of impact include, but are not limited to:
- predicative algorithms for cancer therapy using machine learning,
- molecular dynamics simulation modeling for drug discovery, and
- medical image analysis.
CLICK HERE for the full report
Multivariate Analysis of Transcript Splicing (MATS) is an open tool for transcript slicing that is commonly used by NIH Biowulf users. The MATS package accepts a number of pairs samples of RNA-Seq data to detect differential alternative spicing events. For four samples pairs, the package took around 5 days to generate its results. An NCI investigator wanted to use the package to analyze multiple groups of 32 sample pairs which might take over a month to complete. The HPC/DM group was approached to enable such analysis.
In this blog, we will show how we were able to get an average of 4x speedup in the total runtime of the MATS package which we pushed on Biowulf under the name MATS-NIH.
CLICK HERE for the full report.
