Vertically Integrated Projects: High Performance Computing
A unique undergraduate level project-based course that enables students to work on research projects for more than just 1 semester.
Prof. Chandrasekaran has been teaching the VIP course since Spring 2017 to date and has worked with over 80 VIP students till date.
A VIP-HPC Sample Project
A VIP project started in Spring 2017, under the supervision of Prof. Chandrasekaran, resulted in a publication in one of the top computational biology journals PLOS under Software Collection.
Collaborators: This is work in collaboration with Prof. Juan Perilla and Alex Bryer, a PhD student, from the Department of Chemistry and Biochemistry at the University of Delaware.
Project Team: Eric Wright and Mauricio Ferrato, VIP HPC students, joined this project during their junior year. Upon graduation, they joined my research lab
Robert Searles, my former PhD student, now in NVIDIA, and Alex Bryer, Prof. Perilla’s PhD student served as the senior mentors on the project.
Journal publication DOI: https://journals.plos.org/ploscompbiol/article?id=10.1371/journal.pcbi.1007877
Project Title: Accelerating Prediction of Chemical Shift of Protein Structures on GPUs using OpenACC
Project Abstract: Experimental chemical shifts (CS) from solution and solid state magic-angle-spinning nuclear magnetic resonance (NMR) spectra provide atomic level information for each amino acid within a protein or protein complex. However, structure determination of large complexes and assemblies based on NMR data alone remains challenging due to the complexity of the calculations. Here, we present a hardware accelerated strategy for the estimation of NMR chemical-shifts of large macromolecular complexes based on the previously published PPM_One software. The original code was not viable for computing large complexes, with our largest dataset taking approximately 14 hours to complete. Our results show that serial code refactoring and parallel acceleration brought down the time taken of the software running on an NVIDIA Volta 100 (V100) Graphic Processing Unit (GPU) to 46.71 seconds for our largest dataset of 11.3 million atoms. We use OpenACC, a directive-based programming model for porting the application to a heterogeneous system consisting of x86 processors and NVIDIA GPUs. Finally, we demonstrate the feasibility of our approach in systems of increasing complexity ranging from 100K to 11.3M atoms.
Awards: * Won one of the best research posters in the International Supercomputing Conference (ISC), 2019, Frankfurt, Germany under the HPC category. * Won the best poster in the VIP Mid-Altantic poster competition, Spring 2018. Prof. Chandrasekaran and team have presented this work at GTC 2019, SIAM CSE 2019 and PASC 2019.
Other VIP Projects include
Exploring acceleration of PhysiCell - a bio physics code on GPUs
Building roofline performance models for AMD and NVIDIA GPUs using real-world applications
Submitting kit reports for Standard Performance Evaluation Corporation (SPEC) High Performance Group (HPG’s) suites
Building a database to rank Standard Performance Evaluation Corporation (SPEC) High Performance Group (HPG’s) results using the work in progress HPC2020 benchmark suite.
Updating the OpenACC Best Practices Guide, work in collaboration with NVIDIA
Parallelizing NASA parallel benchmarks (NPB) on multiple platforms thus learning parallel programming concepts
Learning to write scripts to run a large number of bioinformatics applications on large clusters