Minimizing Data Movement in Large Scale Multicores Graph analytics has gained prominence in the era of big data where many applications work on unstructured data. Efficient cache utilization is critical to the performance of graph algorithms. Unlike typical algorithms graph algorithms are characterized by random memory access patterns due to their inherent unstructured form and sparsity. Random accesses patterns cause poor on-chip cache utilization and long latency accesses to DRAM. We will study hardware and software mechanisms to achieve better cache utilization and graph algorithm performance.
I am participating in SuperUROP because I want to gain high level research experience in my major. I took 6.172 (Performance Engineering of Software Systems) last year and I enjoyed finding and fixing performance bottlenecks. I'm interested in computer architecture and I enjoy parallel computing problems. I'm excited to apply my knowledge and interests to my project.