[Bug Report] Severe VRAM Allocation Instability in PyTorch after llama-cpp-python is Imported
I maintainer di solito rispondono entro 1 giorno
Nessuno ha ancora preso questa issue.
Valutazione
- Difficoltà
- 5/5
- Tempo stimato
- Più di una settimana
- Idoneità per principianti
- 30/100
- Tipo di issue
- Bug
- Chiarezza
- Abbastanza chiara
- Stato di attività
- Ferma
- Ambito
- ai-infra-agents
Direzione di ricerca
L’issue descrive un conflitto di basso livello tra i contesti CUDA di llama-cpp-python e PyTorch. Inizia esaminando il codice di inizializzazione CUDA nei bindings di llama-cpp-python, in particolare intorno al momento dell’importazione. Cerca qualsiasi creazione di contesti CUDA o impostazione dell’allocatore di memoria. Riproduci il problema in un ambiente minimale che importi llama-cpp-python e poi carichi un modello PyTorch di grandi dimensioni, monitorando i pattern di allocazione della VRAM. “Done” consisterebbe nell’identificare la chiamata o la configurazione specifica che causa il conflitto e nel proporre una correzione.
Scritto dal modello di indicizzazione a partire dal testo della issue.
Descrizione
Prerequisites
- I am running the latest code.
- I carefully followed the README.md.
- I searched using keywords relevant to my issue to make sure that I am creating a new issue that is not already open (or closed).
- I reviewed the Discussions, and have a new bug to share.
Expected Behavior
When loading a large PyTorch-based model (e.g., an SDXL checkpoint >10GB) within a host application (ComfyUI), I expect the GPU's dedicated memory (VRAM) usage to increase smoothly and stabilize at the required level for the model. This process should be fast and efficient.
Current Behavior
The VRAM allocation is extremely erratic and inefficient. The memory usage graph shows a severe sawtooth/spiky pattern—it spikes up, drops slightly, spikes again, and repeats this cycle until the model is finally loaded. This process is significantly slower than normal and suggests a fundamental conflict in memory management at the CUDA level.
This behavior only occurs when a component that imports llama-cpp-python is loaded by the host application at startup. The issue manifests before any llama-cpp-python functions are actively called; the mere act of importing the library is enough to trigger this conflict with subsequent PyTorch operations.
Environment and Context
-
Hardware:
- GPU: NVIDIA RTX PRO 6000 Blackwell Workstation Edition (94.5 GB VRAM)
-
Operating System:
- Windows 11
-
SDK and Software Versions:
- Python Version:
3.12.11 - Compiler: MSVC / Visual Studio
2022used to compile for CUDA - CUDA Toolkit: 12.8 (locally compiled)
- NVIDIA Driver Version:
580.97 - Host Application: ComfyUI (Release
2025-08-27or newer)
- Python Version:
-
Key Python Packages:
- llama-cpp-python:
0.3.16 - torch:
2.8.0+cu128 - diskcache: 5.6.3
- llama-cpp-python:
Failure Information (for bugs)
This report details a deep-seated conflict between llama-cpp-python and torch regarding CUDA context and/or VRAM management.
Investigation Summary & Key Findings
We conducted an extensive, multi-day investigation to isolate this issue, and the results were conclusive:
-
The problem is triggered by the
importofllama-cpp-python. Any host application that loads this library at startup will cause subsequent, unrelatedtorchVRAM allocations to become unstable. Disabling the components that importllama-cpp-pythoncompletely resolves the issue. -
The
diskcachedependency is a Red Herring. We initially suspected thediskcachelibrary was the cause. To test this, we performed a definitive experiment: we hid the realdiskcachelibrary and replaced it with a completely inert, non-functional fake Python package that only contained an emptyCacheclass.- Result: The VRAM instability problem persisted even with the fake library.
- Log Confirmation: The application log confirmed our fake library was being used:
--- 警告:正在使用一個偽造的 (FAKE) diskcache 函式庫 --- Set up system temp directory: C:\Users\Win\AppData\Local\Temp\latentsync_16880d62 - Conclusion: This proves that
diskcache's code is not the active cause. The problem isllama-cpp-pythonitself. The initial correlation was due todiskcachebeing a mandatory dependency; uninstalling it simply preventedllama-cpp-pythonfrom being imported successfully.
Final Hypothesis
The evidence overwhelmingly suggests a low-level conflict between how llama-cpp-python (especially a version compiled against a modern CUDA 12.8 toolkit) initializes its CUDA context and how PyTorch expects to manage the VRAM. The initialization of llama-cpp-python appears to "pollute" or "fragment" the VRAM state, making subsequent large-scale allocations by PyTorch highly inefficient.
Steps to Reproduce
- Set up a ComfyUI environment with PyTorch and the system components listed above.
- Compile and install
llama-cpp-pythonwith CUDA 12.8 support. - Install any ComfyUI custom node that imports
llama-cpp-pythonat startup (e.g.,ComfyUI-JoyCaptionorComfyUI-Searge-LLM-Node). - Launch ComfyUI. The server will start, and
llama-cpp-pythonwill be imported into the process. - Load a standard workflow designed to load a large PyTorch-based checkpoint model (e.g., an SDXL checkpoint, >10GB).
- Observe the "Dedicated GPU Memory" graph in Windows Task Manager. You will see the erratic, spiky allocation pattern instead of a smooth ramp-up.
Note on llama.cpp comparison: This issue is not about the performance or functionality of llama-cpp-python's inference, but rather a resource conflict it creates with another major library (torch). Therefore, comparing behavior against llama.cpp's ./main executable is not applicable, as the issue requires both runtimes to be active in the same Python process.
- Lingua principale
- Python
- Stelle
- 10.7k
- Fork
- 1.5k
- Merge medio
- 1h 40m
- PR unite (30g)
- 5
Preparare l'ambiente
- Nessun Dockerfile né file Docker Compose
- Nessun modello di pull request
- Leggi la guida per i contributori
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 abetlen/llama-cpp-python
-
Difficoltà 2/5 1-3 ore Idoneità per principianti 78/100
abetlen/llama-cpp-python#2424 ·
I maintainer di solito rispondono entro 1 giorno
-
Seven llama_sampler_init_* bindings admit keyword arguments that the ctypes function object silently dropsForse già presa @Belal0066 l’ha presa 59 giorni fa. Aperta
Difficoltà 2/5 1-3 ore Idoneità per principianti 88/100
abetlen/llama-cpp-python#2371 ·
I maintainer di solito rispondono entro 1 giorno
-
uv add llama-cpp-python wheels fails for versions above 0.3.30Forse già presa Una pull request collegata a questa issue è aperta o già unita. Aperta
Difficoltà 2/5 1-3 ore Idoneità per principianti 65/100
abetlen/llama-cpp-python#2352 · 1 commento · 3 reazioni ·
I maintainer di solito rispondono entro 1 giorno
-
Difficoltà 2/5 1-3 ore Idoneità per principianti 75/100
abetlen/llama-cpp-python#2314 ·
I maintainer di solito rispondono entro 1 giorno
-
Llama.embed() calls LlamaBatch.add_sequence with old 3-arg signature; missing logits_arrayForse di nuovo libera @lxcxjxhx l’ha presa 136 giorni fa e non c’è nessuna pull request aperta. Aperta
Difficoltà 2/5 1-3 ore Idoneità per principianti 65/100
abetlen/llama-cpp-python#2211 · 2 commenti ·
I maintainer di solito rispondono entro 1 giorno
Tutte le issue di abetlen/llama-cpp-python
Issue simili
-
Bug IronPython Python 3
Difficoltà 2/5 1-3 ore Idoneità per principianti 72/100
pyrevitlabs/pyRevit#4184 ·
I maintainer di solito rispondono entro 1 giorno
-
Difficoltà 1/5 Meno di un'ora Idoneità per principianti 74/100
jqueryscript/awesome-claude-code#1009 ·
-
tests
Difficoltà 2/5 1-3 ore Idoneità per principianti 66/100
I maintainer di solito rispondono entro 3 giorni
-
Difficoltà 2/5 1-3 ore Idoneità per principianti 72/100
I maintainer di solito rispondono entro 1 giorno
-
[Plugin] 群聊记忆 · 人格养成Aperta
Difficoltà 1/5 1-3 ore Idoneità per principianti 74/100