Joshua Hoke Davis

I am a fourth-year computer science Ph.D. student at the University of Maryland in College Park, MD, where I am advised by Prof. Abhinav Bhatele.

As a member of the Parallel Software and Systems Group, the primary subject of my research as a doctoral student is performance portability, or the ability of a single-source application to achieve good performance on a range of hardware platforms. We are presenting some initial results from this effort, comparing performance portability enabled by GPU programming models, in our SC23 best poster and associated arXiV paper. More information on this work can also be found on the PSSG website.

Earlier in my time at UMD I collaborated with the CRoCCo Lab on adding GPU acceleration and adaptive mesh refinement to their finite-difference computational fluid dynamics code using AMReX. We presented the results of this work along with extensive performance analysis at IPDPS 2023.

I graduated from the University of Delaware in 2021 with a B.S. in Computer Science as well as a B.A. in Philosophy and minor in Mathematics. At UD, I worked with the Computational Research and Programming Lab, led by Prof. Sunita Chandrasekaran, as a developer for the SOLLVE OpenMP V&V test suite, part of the Exascale Computing Project (ECP). To see the latest results from the V&V suite, check out their website.

At the moment I am additionally involved with automatically formally verifying data structures from The Art of Multiprocessor Programming, using CIVL, collaborating with the Verified Software Lab at UD, led by Prof. Stephen Siegel. More on that project can be found here on GitHub.

My CV is available here.