High-Performance Application Studies

I worked with Abdul Dakkak and Cheng Li at Illinois on a full-stack performance analasis platform for machine learning software. The same infrastructure is also used to deliver some course projects for GPU programming courses.

I worked with Mert Hidayetoglu at Illinois on GPU acceleration of a massively-parallel tomographic image reconstruction code.

I was part of the IME-Accelerators team for the NCSA PAID program. This program pairs domain scientists with computing specialists to improve effective utilization of the Blue Waters supercomputer at the National Petascale Computing Facility. I have worked on:

  • CUDA Kernel Optimization for AWP with the Southern California Earthquake Center
  • OpenACC Acceleration for SETSM with Ohio State University
Avatar
Carl Pearson
Electrical and Computer Engineering Ph.D. Candidate

PhD Candidate

Publications

A Fast and Massively-Parallel Solver for Multiple-Scattering Tomographic Image Reconstruction

We present a massively-parallel solver for large Helmholtz-type inverse scattering problems. The solver employs the distorted Born …

Comparative Performance Evaluation of Multi-GPU MLFMM Implementation for 2-D VIE Problems

We compare multi-GPU performance of the multilevel fast multipole method (MLFMM) on two different systems: A shared-memory IBM S822LC …

Scalable Parallel DBIM Solutions of Inverse-Scattering Problems

We report scalable solutions of inverse-scattering problems with the distorted Born iterative method (DBIM) on large number of …

Thoughts on Massively-Parallel Heterogeneous Computing for Solving Large Problems

In this paper, we present our view of massively-parallel heterogeneous computing for solving large scientific problems. We start by …

Large Inverse-Scattering Solutions with DBIM on GPU-Enabled Supercomputers

We report inverse-scattering solutions on supercomputers involving large numbers of graphics processing units (GPUs). The …

Talks

Bigger GPUs and Bigger Nodes

GPU Performance Nuggets

Experience from GPU acceleration of scientific code on NCSA Blue Waters