Idle CPU: a fixed 10 ms main loop, and a PulseAudio stream that stays open after the stream ends
I maintainer di solito rispondono entro 1 giorno
Nessuno ha ancora preso questa issue.
Valutazione
- Difficoltà
- 4/5
- Tempo stimato
- 3-5 giorni
- Idoneità per principianti
- 42/100
- Tipo di issue
- Bug
- Chiarezza
- Abbastanza chiara
- Stato di attività
- Attiva
- Stack tecnologico
- cpp, linux
- Ambito
- audio-video-rtc, performance
Direzione di ricerca
Start with the fixed sleep in the main loop: LOOP_INTERVAL_MS at src/main.cpp:57 and the sleep_for block at src/main.cpp:733-743. The PulseAudio cork/close fix belongs in the code that handles the stream-ended event, which the payload does not name, so locate it first. Run the player with -o pulse, go idle after a stream ends, and confirm with pactl that the sink input is corked and the sink suspends. Done means both idle wakeup rates drop and the sink suspends, but the loop change may need the library to expose a waitable fd first.
Scritto dal modello di indicizzazione a partire dal testo della issue.
Descrizione
Summary
An idle player has a CPU floor that never drops, and a stream that has ended keeps the sound server busy too. Neither is large, but both show up in "why does an idle player use CPU at all" reports. Downstream: music-assistant/local-audio-addon#31, where an idle player used about three times what the Music Assistant server did.
Measured on v0.3.0 (sendspin-cpp v0.8.0), -o pulse into Home Assistant OS's PulseAudio, x86_64, 4 cores.
1. The main loop polls every 10 ms, idle or not
src/main.cpp:57 sets LOOP_INTERVAL_MS = 10. Every pass polls mDNS, the control socket, the network provider, the sink and the hooks, then does sleep_for(10ms) (main.cpp:733-743), whether or not anything is happening.
| Thread | Wakeups/s | CPU |
|---|---|---|
main (sendspin-cli) |
98 | ~0.9% of one core, idle after a stream |
That is the whole idle floor of the process. A loop that blocks on its file descriptors (the control socket, the mDNS fd, the websocket) with a timeout set by the next due timer would bring an idle player close to zero wakeups. The catch is that sendspin-cpp's network provider is polled from this loop (network provider polled only from the main loop in the v0.9.5 notes), so the library would need to expose something to wait on.
2. The native PulseAudio stream is never corked or closed when a stream ends
After I player: Stream ended, with status reporting state: paused and stream: idle, the PulseAudio sink input stays connected and uncorked:
Sink Input #3
Sink: 0
Corked: no
application.name = "sendspin-cli"
0 alsa_output.pci-0000_00_1f.3.analog-stereo ... RUNNING
The result:
- libpulse's I/O thread (
threaded-ml) keeps waking about 97 times a second. - The server's sink stays
RUNNINGand never suspends. The PulseAudio daemon itself sat at 1.6% of a core, almost all in its ALSA sink thread. - Any USB DAC or amp behind the sink is held awake, and module-suspend-on-idle never gets the chance to act.
Corking the stream at stream end (and uncorking at the next start), or closing it, would let the sink suspend. I haven't checked whether the ALSA, PipeWire or PortAudio sinks hold their device the same way.
How it was measured
Per-thread utime+stime deltas from /proc/<pid>/task/*/stat over 10-20 s windows, and voluntary plus nonvoluntary context-switch deltas from /proc/<pid>/task/*/status for the wakeup rates. Sink state is from pactl list sink-inputs and pactl list short sinks inside the Home Assistant audio container.
- Lingua principale
- C++
- Stelle
- 14
- Fork
- 5
- Merge medio
- 1g 12h
- PR unite (30g)
- 26
Preparare l'ambiente
Questo progetto non fornisce container di sviluppo, Dockerfile né guida per i contributori, quindi l'ambiente è a tuo carico: parti dal suo README e consulta la nostra guida al primo contributo per i passaggi generali.
Come iniziare
- Leggi tutta la issue e poi la guida ai contributi del progetto.
- Commenta sulla issue per dire che te ne occupi tu — evita che due persone facciano lo stesso lavoro.
- Fai un fork del repository e lavora su un branch.
- Apri una pull request che faccia riferimento al numero della issue.
Altre issue di Sendspin/sendspin-cpp-cli
-
Difficoltà 2/5 1-3 ore Idoneità per principianti 72/100
Sendspin/sendspin-cpp-cli#111 ·
I maintainer di solito rispondono entro 1 giorno
-
Feature request: keep the ALSA output running briefly between streams (--keep-output-alive <s>)Aperta
Difficoltà 4/5 3-5 giorni Idoneità per principianti 42/100
Sendspin/sendspin-cpp-cli#109 · 1 commento ·
I maintainer di solito rispondono entro 1 giorno
-
Difficoltà 3/5 1-2 giorni Idoneità per principianti 35/100
Sendspin/sendspin-cpp-cli#108 ·
I maintainer di solito rispondono entro 1 giorno
-
Difficoltà 3/5 1-2 giorni Idoneità per principianti 62/100
Sendspin/sendspin-cpp-cli#106 ·
I maintainer di solito rispondono entro 1 giorno
-
Difficoltà 3/5 1-2 giorni Idoneità per principianti 35/100
Sendspin/sendspin-cpp-cli#97 · 1 commento ·
I maintainer di solito rispondono entro 1 giorno
Tutte le issue di Sendspin/sendspin-cpp-cli
Issue simili
-
bug
Difficoltà 2/5 1-3 ore Idoneità per principianti 62/100
pocoproject/poco#5626 ·
I maintainer di solito rispondono entro 2 giorni
-
Difficoltà 2/5 1-3 ore Idoneità per principianti 78/100
-
Issue-Bug Needs-Triage
Difficoltà 2/5 1-3 ore Idoneità per principianti 62/100
microsoft/winget-cli#6645 ·
I maintainer di solito rispondono entro 2 giorni
-
cleanup help wanted
Difficoltà 2/5 1-3 ore Idoneità per principianti 62/100
ESPresense/ESPresense#2602 ·
I maintainer di solito rispondono entro 1 giorno
-
devex good first issue
Difficoltà 2/5 1-3 ore Idoneità per principianti 62/100
I maintainer di solito rispondono entro 1 giorno