Interested in attending? Contact events [at] cs.stonybrook.edu" style="text-decoration-line: none; color: rgb(34, 34, 34); font-family: Effra-Rg, Verdana, Arial, Helvetica, sans-serif; cursor: default; white-space: pre-wrap;">events [at] cs.stonybrook.edu for the Zoom info.  

Title: Algorithmic foundation of parallel paging and scheduling under memory constraints



Abstract: This thesis develops an algorithmic foundation for paging and space-constrained computation in shared-memory systems. In particular, we present tight upper and lower bounds on the competitive ratios for parallel paging in multicore shared memory. We also extend the theory to manage modern-day memory systems called High-Bandwidth Memory (HBM), common to new supercomputers. We show how the parallel running time of a program depends on the available memory and present a provably efficient way to achieve near-optimal parallel running time for a given amount of memory.