Hello World! I'm David Krasowska.
david:~$ whoami
David Krasowska
Ph.D. Candidate at Northwestern
Northwestern University (2023 – Now)
Ph.D. Candidate & M.S. in Computer Science
Clemson University (2019 – 2022)
B.S. in Computer Engineering
david:~$ summary
david:~$ education
david:~$ contact
David Krasowska is a Ph.D. candidate at Northwestern University, advised by Dr. Peter Dinda. His research began at Clemson University collaborating with Argonne National Laboratory on lossy compression for HPC scientific applications. He received the DOE Computational Science Graduate Fellowship for his graduate studies. Currently he explores raising the level of abstraction for programming emerging accelerators, such as multi-GPU systems and Processing-in-Memory (PIM).
Northwestern University (2023 – Now)
Ph.D. Candidate & M.S. in Computer Science
Clemson University (2019 – 2022)
B.S. in Computer Engineering
Experience
Jun 2024
Aug 2024
Aug 2024
Visiting Student
Los Alamos National Laboratory
Task based heterogeneous runtime research for UPMEM processing in memory (PIM). Adding backend support for UPMEM PIM into Legion runtime programming system (Github). Presented current state of the work at the Legion Retreat.
Jun 2022
Jun 2023
Jun 2023
Research Aide
Argonne National Laboratory
Continuation of prior work at Clemson University to create black-box regression model with 3D dataset capabilities. Led to multiple publications and awards. Contributed to Libpressio, an Argonne library for compression.
Jun 2021
Dec 2022
Dec 2022
High Performance Computing Creative Inquiry
Clemson University
Participant in the Student Cluster Competition (SCC) at SC '21 and INDY SCC at SC '22. Collaboration with Dell and Intel to build a cluster with greatest performance per watt. Set up a distributed cluster with package managers, applications, and benchmarks. Gained knowledge of parallel computing with MPI.
May 2021
May 2022
May 2022
Undergraduate Student Researcher
Clemson University, FTHPC
Lossy compression research with collaboration Argonne National Laboratory and Clemson University FTHPC using the Palmetto Cluster. Analyzing statistical correlations within datasets in comparison to compression performance. First ever publication: DRBSD-7.
Publications and Technical Reports
2026
Easy, Yet Performant PIM Through Deferred Execution That Enables Vertical and Horizontal Fusion
D. Krasowska, A. Crotty, P. Dinda
submitted to ASPLOS; awaiting decision
2026
cuNumeric.jl: Automated Performant Multi-GPU Numerical Computing for Julia
D. Krasowska, E. Meitz, W. Lee, P. Dinda
submitted to IPDPS; awaiting decision
2025
K. Nagaitsev, G. Dube, K. Hallsby, P. Liu, Q. Jiang, L. Myers, D. Krasowska, A. Butler, R. Xu, P. Dinda
Technical Report NU-CS-2025-33, Northwestern University
2023
R. Underwood, J. Bessac, D. Krasowska, J. C. Calhoun, S. Di, F. Cappello
International Journal of High Performance Computing Applications (IJHPCA)
2023
A. Ganguli, R. Underwood, J. Bessac, D. Krasowska, J. C. Calhoun, S. Di, F. Cappello
IEEE International Conference on Cluster Computing (CLUSTER)
2021
D. Krasowska, J. Bessac, R. Underwood, J. C. Calhoun, S. Di, F. Cappello
7th Workshop on Data Analysis and Reduction for Big Scientific Data (DRBSD-7)
MPI
CUDA
C
C++
Python
Julia
LLVM
HPC