Coarse-grain speculation for emerging processors
Hari K. Pyla · 2011
The impending multi/many-core processor revolution requires that programmers leverage explicit concurrency to improve performance. Unfortunately, a large body of applications/algorithms are inherently hard to parallelize due to execution order constraints imposed by data and control dependencies or being sensitive to their input data and not scale perfectly, leaving several cores idle. The goal of this research is to enable such applications leverage multi/many-cores efficiently to improve their performance.