Lines Matching refs:as

5 …as interactive threads and low priority threads as batch threads. It then uses a timesharing model…
10 … clutch timesharing per cluster and adds other multi-cluster features such as thread placement, mi…
14 …l decisions at various thread grouping levels. The hierarchical scheduler, as its implemented toda…
22 …t are represented by a single entry at this level. These entries are known as *root buckets* throu…
26 … next for execution. Each root bucket with runnable threads is represented as an entry in a priori…
53 …as described above. The implementation allows extremely low latency CPU access for Above UI thread…
55 …adline ordering, the root bucket selection algorithm marks the root bucket as being in starvation …
72 …ch root hierarchy. Each clutch bucket with runnable threads is represented as an entry in a runque…
75 … of voluntary blocking time and CPU usage time for the clutch bucket group as a whole. This score …
80 …re pending execution for this clutch bucket group. This value is reset as soon as one of the threa…
87 …ons on this abstraction allows the system to make interesting choices such as preferring Apps over…
93 …as an entry in a runqueue which is organized based on the `sched_pri` of threads. The thread selec…
97 …as they use CPU. The priority decay algorithm attempts to reward bursty interactive threads and pe…
116 …as threads are inserted/removed from the hierarchy. The root level also maintains the urgency bits…
124 …t priority (promoted or base pri) thread within that clutch bucket and assign that as root priority
142 …he highest runnable thread and interactivity score. The actual calculation algorithm is as follows:
164 …ad placement and load balancing, the Edge scheduler represents the machine as a graph where each n…
168 …be **compact**. Limit small-width workload threads to a single cluster as much as possible. A few …
174 …* Efficiency cores offer meaningful performance uplift for benchmarks as well as throughput-orient…
175 …res that the high QoS threads experience low scheduling latency under heavy CPU contention as well.
180 …as thread groups, asymmetric spill & steal, etc. were invented to efficiently exploit this perform…
190 …ommendation of the sched\_clutch\_bucket_group it belongs to and uses that as  the starting decisi…
192 …on to migrate the running and runnable threads for that group immediately (as opposed to the next …
228 Both metrics are maintained as an exponentially moving weighted average to make sure they capture t…
231 …get the scheduling latency metric for the thread's QoS and preferred cluster (as specified by CLPC)
238 *    Otherwise, if delta is greater than largest delta, store delta as largest delta
250 …fies clutch buckets (and correspondingly the threads in the clutch bucket) as foreign when these c…
260 *     If available, return THREAD_NULL for the steal callout and perform rebalancing as part of SCH…