Garima Gupta studied Computer Science at the University of Windsor from 2004 with a specialty in High Performance Computing and received her Master degree in 2006. Under the supervision of Dr. Angela Sodan, Garima worked on performance analysis of multicore/multithreaded processor systems, on parallel computing, and on grid scheduling. Her experimental work was done on a CFI funded cluster of her supervisor and Sharcnet clusters (CFI funded Ontario network of clusters).
Her thesis topic was in the area of multi-dimensionally adaptive grid middleware and explores trade-offs between partitioning data into chunks for easier redistribution or finding an optimal distribution from scratch. It shows benefits from adaptive resource allocation under resource sharing by multiple jobs. The latter permits increasing utilization and exploitation of modern cluster nodes with multiple CPUs, multi-core CPUs, or hyperthreaded CPUs. The results were published at the ACM Workshop of Adaptive and Reflective Middleware (ARM) in Melbourne and the PDCS conference in Dallas. An extended version was recently accepted for the Journal of Future Generation Computer Systems, Elsevier.
Garima’s dedicated work has landed her a job at Platform Computing (Toronto) which is an internationally well-known company in grid computing.