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

Tagged “software performance engineering”