Tags

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