Home
last modified time | relevance | path

Searched refs:metrics (Results 1 – 9 of 9) sorted by relevance

/xnu-12377.61.12/doc/observability/
H A Drecount.md11 These two metrics are abbreviated to cycles-per-instruction, or CPI, for brevity.
53 Tables also include formulaic columns that can be derived from two metrics, like CPI or power.
79 - `ru_metrics[RCT_LVL_COUNT]`: metrics accumulated in each exception level
82 The metrics are stored in a `recount_metrics` structure with the following fields:
93 During preemption, the context switch handler attributes metrics back to the level stored in each t…
94 …d `recount_leave_secure` update the current thread's level and attribute metrics back to the previ…
105 To ensure the accuracy of formulas involving multiple metrics, like CPI, all metrics must be update…
111 The sequence lock doesn't need to be taken for these metrics since they're never updated simultaneo…
121 Unlike the other metrics, the energy counters are not sampled directly by Recount so the values can…
H A Dcoalitions.md39 A ledger is a lightweight key-value store that we use to track various metrics. Ledgers are created…
43 …y B, a reader thread might only see the update to A. Therefore computing metrics from multiple ent…
/xnu-12377.61.12/tools/lldbmacros/
H A Drecount.py110 metrics = usage.ru_metrics[level]
111 self._times_mach[level] += unsigned(metrics.rm_time_mach)
112 if hasattr(metrics, "rm_cycles"):
113 self._instructions[level] += unsigned(metrics.rm_instructions)
114 self._cycles[level] += unsigned(metrics.rm_cycles)
115 if unsigned(metrics.rm_cycles) != 0:
280 metrics = usage.ru_metrics[level]
281 time = unsigned(metrics.rm_time_mach)
287 if hasattr(metrics, "rm_cycles"):
288 cycles = unsigned(metrics.rm_cycles)
[all …]
/xnu-12377.61.12/tests/recount/
H A Drecount_perf_tests.c87 metrics_t metrics = s->us_metrics; in interface_perf_test() local
91 metrics &= ~(METRIC_INSNS | METRIC_CYCLES); in interface_perf_test()
95 for (int metric = next_bit(&metrics); metric >= 0; in interface_perf_test()
96 metric = next_bit(&metrics)) { in interface_perf_test()
/xnu-12377.61.12/osfmk/kern/
H A Drecount.c340 struct recount_metrics *metrics = &usage->ru_metrics[level]; in recount_usage_add_snap() local
342 metrics->rm_time_mach += snap->rsn_time_mach; in recount_usage_add_snap()
344 metrics->rm_cycles += snap->rsn_cycles; in recount_usage_add_snap()
345 metrics->rm_instructions += snap->rsn_insns; in recount_usage_add_snap()
/xnu-12377.61.12/doc/debugging/
H A Dextensible_paniclog.md213 We want to add the utilization metrics of the panic log to the panic.ips file. This will give us an…
253 …hese calculations will use the data that we collect from our utilization metrics and leave out spa…
/xnu-12377.61.12/bsd/kern/code_signing/
H A Dtxm.c611 txm_metrics = &txm_rw_data->metrics; in get_code_signing_info()
1863 SYSCTL_NODE(_txm, OID_AUTO, metrics, CTLFLAG_RD, 0, "TXM Metrics");
/xnu-12377.61.12/doc/scheduler/
H A Dsched_clutch_edge.md79 The clutch bucket group maintains a few metrics to allow calculation of the interactivity score for…
230 Both metrics are maintained as an exponentially moving weighted average to make sure they capture t…
/xnu-12377.61.12/bsd/net/
H A Dntstat.c483 SYSCTL_PROC(_net_stats, OID_AUTO, metrics,
833 struct nstat_metrics *metrics = &client_info->nstat_metrics; in metrics_for_client_id() local
838 *metrics = nstat_metrics; in metrics_for_client_id()
843 *metrics = nstat_metrics; in metrics_for_client_id()
846 nstat_accumulate_client_metrics(metrics, client); in metrics_for_client_id()
857 *metrics = client->ntc_metrics; in metrics_for_client_id()