Tags

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