scx_flow stats

per-CPU queues, fixed slices and placement

Loading...

strict
● running

TL;DR

Loading...

Explanation

E_saved = (P_perfP_strict) × t ÷ 1000

waiting for 3 accepted pairs

E_saved is saved energy in kWh over the span.

P_perf is mean perf power in W over kept seconds in all accepted pairs.

P_strict is mean strict power in W over kept seconds in all accepted pairs.

t is hours in the span.

unavailable

These figures are estimates based on real measurements collected while scx_flow is running. They approximate the energy saved by strict mode since launch. For a more exact comparison, record energy use with scx_flow in strict mode and without scx_flow, then compare the two readings.

Light depth: 0
Hog depth: 0
Allowance: 0
Demote rate: 0
Promote rate: 0
Wake rate: 0
Steal rate: 0
Skip rate: 0
Preempt rate: 0
Preempt skip (all): 0
Skip armed: 0
Skip deserved: 0
Skip group: 0
Skip mask: 0
Skip rate: 0
Coalesce rate: 0
Slot rate: 0
On-CPU: 0
Uptime: 0
Runtime: 0
Inserts: 0
Requeues: 0
Completions: 0
Park moves: 0
Steal moves: 0
Steal cross: 0
Kicks: 0
No task state: 0
EDF enqueued: 0
EDF clamped: 0
EDF ordered: 0
Demote: 0
Promote: 0
Wake promote: 0
Pinned inflate: 0
Group skip: 0
Preempt kicks: 0
Preempt skip (all): 0
Skip armed: 0
Skip deserved: 0
Skip group: 0
Skip mask: 0
Skip rate: 0
Kick coalesced: 0
Slot moves: 0
Slot defer: 0
Slot kicks: 0
Wheel skips (frozen): 0

Loading...