Tags

instructors seminar atomic get_involved cilk Chip multiprocessor cilk_for cilk_scope cilk_spawn core cpu commutative operation deadlock data race cilk_sync concurrent agent distributed memory critical section determincacy race execution time span fake lock false sharing hyperobject instruction knot lock lock contention nondeterminism linear speedup determinism mutex multiprocessor nonlocal variable multicore parallel loop perfect linear speedup opencilk race condition parallelism processor receiver process running time serial execution response time sequential consistency scale up scale down reducer global variable serial semantics shared memory software performance engineering serial projection strand sync spawn speedup thread throughput task-parallel worker work stealing events work competition view Cilksan