Email: mbrodowicz@tacc.utexas.edu
Maciej Brodowicz has joined TACC in 2025 to conduct research in high performance computing technologies with specific focus on large scale dynamic and irregular problems, including graph computing. Among other approaches, he investigates scalable, power-efficient, memory-centric, fine grain cellular architecture known as the Active Memory Architecture (AMA). Besides machine functions and communication protocol specification, this involves optimized ISA encoding with compiler support, full HDL hardware implementation, and performance simulation for select applications. This work received several funding grants, including from NASA and IARPA.
Previously, Maciej held research positions at University of Houston, California Institute of Technology, Louisiana State University, and Indiana University. In these institutions he investigated scalable I/O algorithms to support checkpointing of large scale scientific simulations, developed object-oriented parallel file system to support genomic computations, produced parametric execution models that resulted in creation of the HPX family of low-overhead runtime systems for efficient asynchronous computations on distributed memory computers, designed and implemented FPGA logic for algorithm acceleration, developed compiler plugins for automated code transformations, and build multiple tracing and performance monitoring tools to aid analysis of application behavior on supercomputers.
B. Chandio, M. Brodowicz, and T. Sterling. Exploring the Design Space for Message-Driven Systems for Dynamic Graph Processing Using CCA. Accepted to PPAM ’24, 2024,
T. Sterling, M. Anderson, and M. Brodowicz. High Performance Computing: Modern Systems and Practices, 2nd ed. Morgan Kaufmann, 2024,
T. Sterling, M. Brodowicz, and M. Anderson. First Draft of a Report on the Continuum Computer Architecture. Advances in Parallel Computing, vol. 34, 183-205, 2019,
T. Sterling, M. Anderson, and M. Brodowicz. A Survey: Runtime Systems for High Performance Computing. Supercomputing Frontiers and Innovations, 4(1), 48-68, 2017,
T. Sterling, M. Brodowicz, D. Kogler, and M. Anderson. Asymptotic Computing – Undoing the Damage. New Frontiers in High Performance Computing and Big Data, IOS Press, vol. 30, 55-73, 2017,
T. Sterling, D. Kogler, M. Anderson, and M. Brodowicz. SLOWER: A Performance Model for Exascale Computing. Supercomputing Frontiers and Innovations, 1(2), 42-57, 2014,
S. Yang, M. Brodowicz, W. B. Ligon, and H. Kaiser. PXFS: A Persistent Storage Model for Extreme Scale. Proceedings of ICNC ’14, 2014,
M. Anderson, M. Brodowicz, A. Kulkarni, and T. Sterling. Performance Modeling of Gyrokinetic Toroidal Simulations for a Many-tasking Runtime System. Proceedings of Supercomputing 2013, 136-157, 2013,
M. Anderson, M. Brodowicz, T. Sterling, H. Kaiser, and B. Adelstein-Lelbach. Tabulated Equations of State with a Many-tasking Execution Model. IEEE International Symposium on Parallel & Distributed Processing, 1691-1699, 2013,
T. Sterling, M. Brodowicz, and T. Gilmanov. Towards Brain-inspired System Architectures. BrainComp 2013, Lecture Notes in Computer Science, vol. 8603, 159-170, 2013,
C. J. Michael, M. Brodowicz, and T. Sterling. Improving Code Compression Using Clustered Modalities. Proceedings of the ACM Southeast Regional Conference, 423-428, 2008,
W. Benger, S. Venkataraman, A. Long, G. Allen, S. D. Beck, M. Brodowicz, J. MacLaren, and E. Seidel. Visualizing Katrina – Merging Computer Simulations with Observations. PARA 2006, Springer Lecture Notes in Computer Science vol. 4699, 340-350, 2007.
Design of custom Active Memory Architecture for accelerated processing of dynamic graph problems;
Development of C/C++ compiler back-end supporting the AMA target using the LLVM framework.
