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

Tagged “software performance engineering”