Tags

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