Evaluating work distribution patterns for parallel bitmap compression over SMPs
Ben McCamish, Xinghui Zhao, David Chiu, Jason Sawin, Guadalupe M. Canahuate · 2015
Bitmap indices are popular in managing large-scale data, but their size quickly grows out-of-core without compression. At the same time, Moore's enables a proliferation of machines with parallel architectures, letting users exploit symmetric multiprocessors (SMP) for common tasks. In this poster, we evaluate two widely used parallel work distribution models for parallelizing bitmap compression.