Hacktoberfest 2026: le issue che i maintainer hanno segnato per ottobre, aperte e adatte ai principianti. Sfoglia le issue Hacktoberfest

[bug]: [v7] Idle invokeai-web pins one CPU core at 100% (7.0.0a2-rocm, Radeon AI PRO R9700)

Aperta
  • Qualcosa non va?
#9,998 5 commenti 1 reazione 0 assegnatari Vedi su GitHub

I maintainer di solito rispondono entro 2 giorni

Nessuno ha ancora preso questa issue.

Valutazione

Difficoltà
4/5
Tempo stimato
3-5 giorni
Idoneità per principianti
48/100
Tipo di issue
Bug
Chiarezza
Abbastanza chiara
Stato di attività
Attiva
Stack tecnologico
python
Ambito
backend, performance

Direzione di ricerca

Start by comparing the startup and idle behavior in 7.0.0a2-rocm with 6.14.0-rocm, then trace what creates or uses the default asyncio executor thread. The report identifies the idle thread as _worker in concurrent.futures.thread, but does not identify the application entry point or the cause. Done means an idle instance no longer pins a CPU core while preserving normal startup and generation behavior.

Scritto dal modello di indicizzazione a partire dal testo della issue.

Descrizione

bug
Is there an existing issue for this problem?
  • I have searched the existing issues
Install method

Manual

Operating system

Linux

GPU vendor

AMD (ROCm)

GPU model

Radeon AI PRO R9700

GPU VRAM

32

Version number

7.0.0a2-rocm

Browser

Chrome (Brave)

System Information

Version: 7.0.0-alpha.2, Docker image ghcr.io/invoke-ai/invokeai:7.0.0a2-rocm
GPU: AMD Radeon AI PRO R9700 (gfx1201), 32 GB VRAM
Host: unRAID, Ryzen 7 3800X, 32 GB RAM

What happens:
With no queue running and no browser open, invokeai-web uses 100% of one
CPU core permanently, starting right after startup.
6.14.0-rocm on the same machine idles at ~0%.

Reproduced with an empty root directory (no models, fresh DB):
docker run --rm -d --device /dev/kfd --device /dev/dri
-e INVOKEAI_ROOT=/invokeai_root -v /tmp/invtest:/invokeai_root
ghcr.io/invoke-ai/invokeai:7.0.0a2-rocm
-> docker stats: ~100% CPU

Diagnosis:

  • One thread busy: the default asyncio executor thread "asyncio_0".
  • py-spy (1000 samples / 10 s): 999 samples in
    _worker (concurrent/futures/thread.py:90), no work item, no run() frame,
    i.e. the thread is idle waiting for work.
  • /proc//task//syscall reports "running" on every sample,
    so the wait never blocks in the kernel.
  • py-spy --native fails here (UNW_EBADREG).

Possibly related: vLLM reports the same idle 1-core spin on the
Radeon AI Pro 9700 with ROCm 7.2.x.

What happened

Version: 7.0.0-alpha.2, Docker image ghcr.io/invoke-ai/invokeai:7.0.0a2-rocm
GPU: AMD Radeon AI PRO R9700 (gfx1201), 32 GB VRAM
Host: unRAID, Ryzen 7 3800X, 32 GB RAM

What happens:
With no queue running and no browser open, invokeai-web uses 100% of one
CPU core permanently, starting right after startup.
6.14.0-rocm on the same machine idles at ~0%.

Reproduced with an empty root directory (no models, fresh DB):
docker run --rm -d --device /dev/kfd --device /dev/dri
-e INVOKEAI_ROOT=/invokeai_root -v /tmp/invtest:/invokeai_root
ghcr.io/invoke-ai/invokeai:7.0.0a2-rocm
-> docker stats: ~100% CPU

Diagnosis:

  • One thread busy: the default asyncio executor thread "asyncio_0".
  • py-spy (1000 samples / 10 s): 999 samples in
    _worker (concurrent/futures/thread.py:90), no work item, no run() frame,
    i.e. the thread is idle waiting for work.
  • /proc//task//syscall reports "running" on every sample,
    so the wait never blocks in the kernel.
  • py-spy --native fails here (UNW_EBADREG).

Possibly related: vLLM reports the same idle 1-core spin on the
Radeon AI Pro 9700 with ROCm 7.2.x.

What you expected to happen

Same appr. idle load as under 6.1x

How to reproduce the problem

Turn on the Docker Container with InvokeAI

Additional context

Constant issue. Generating images or not makes no difference.

Discord username

C_Dollerup

Lingua principale
TypeScript
Stelle
28.8k
Fork
3.1k
Merge medio
3g 2h
PR unite (30g)
159

Preparare l'ambiente

  • Nessun Dockerfile né file Docker Compose
  • Ha un modello di pull request
  • Nessuna guida per i contributori

Come iniziare

  1. Leggi tutta la issue e poi la guida ai contributi del progetto.
  2. Commenta sulla issue per dire che te ne occupi tu — evita che due persone facciano lo stesso lavoro.
  3. Fai un fork del repository e lavora su un branch.
  4. Apri una pull request che faccia riferimento al numero della issue.

Altre issue di invoke-ai/InvokeAI

Tutte le issue di invoke-ai/InvokeAI

Issue simili

Altre issue su TypeScript

Ricevi le nuove issue nella tua casella

Un breve riepilogo di issue GitHub adatte ai principianti.