{"id":893,"date":"2023-08-15T11:16:38","date_gmt":"2023-08-15T18:16:38","guid":{"rendered":"https:\/\/labs.wsu.edu\/scads\/?page_id=893"},"modified":"2025-08-20T16:54:24","modified_gmt":"2025-08-20T23:54:24","slug":"parallel-computing","status":"publish","type":"page","link":"https:\/\/labs.wsu.edu\/scads\/research\/parallel-computing\/","title":{"rendered":"Parallel Computing"},"content":{"rendered":"\n

Parallel Computing<\/h1>\n\n\n\n
\n\n\n\n

Graph (network) computations are critical kernels in many algorithms in data mining, data analysis, scientific computing, computational science and engineering, etc. In large-scale applications, the graph computations need to be performed in parallel. Parallelizing graph algorithms effectively \u2014 with emphasis on scalability and performance \u2014 is particularly challenging for a variety of reasons: In many graph algorithms runtime is dominated by memory latency rather than processor speed, there exist little computation to hide memory access costs, data locality is poor, and available concurrency is low.<\/p>\n\n\n\n

Listed below in reverse chronological order are papers we have written together with a number of different collaborators introducing a range of techniques for dealing with these challenges in the context of a variety graph problems. Much of Assefaw’s earlier work in this direction focused on distributed-memory platforms. His more recent effort targets the emerging and rapidly growing multicore platforms as well as massively multithreaded platforms. The list also includes his recent works on other combinatorial problems than graph problems and on problems around matrix computations. At SCADS, we are in general interested in exploring the interplay between algorithms, architectures, and applications in developing scalable systems.<\/p>\n\n\n\n

\n\n\n\n

Software<\/h2>\n\n\n\n

MPI-based implementations of some of the distributed memory parallel algorithms for the distance-1 and distance-2 coloring problems we developed together with colleagues are publicly available via the Zoltan<\/a> software toolkit of Sandia National Laboratories.<\/p>\n\n\n\n

\n\n\n\n

Papers<\/h2>\n\n\n
\r\n \n
\r\n\t\n
\n\t\n\t\t\t\n\t\tDistributed memory architectures\t\t\t<\/a>\n<\/div>\n\n<\/div>\r\n\n\n
\r\n\t\n
\n\t\n\t\t\t\n\t\tMulti-core and multi-threaded architectures\t\t\t<\/a>\n<\/div>\n\n<\/div>\r\n\n\n
\r\n\t\n
\n\t\n\t\t\t\n\t\tShared memory architectures\t\t\t<\/a>\n<\/div>\n\n<\/div>\r\n\n<\/div>\n\n\n
\n\n\n\n

Distributed-memory architectures <\/strong><\/h3>\n\n\n\n