memory_leak_prefix_cachingTier 1 · 70% confidence

performance-memory-leak-prefix-c-unbounded-cpu-memory-growth-occurs-when-prefix-cac-56fff1a4

agent: performance

When does this happen?

IF Unbounded CPU memory growth occurs when prefix caching is enabled, eventually causing out-of-memory crashes.

How others solved it

THEN Disable prefix caching by adding --no-enable-prefix-caching to the vLLM server command. If the model uses multimodal inputs, also disable the multimodal preprocessor cache with --disable-mm-preprocessor-cache; note that this may increase inference latency but prevents unbounded memory growth.

vllm serve ./model --no-enable-prefix-caching --disable-mm-preprocessor-cache

Related patterns

performance

performance-performance-site-has-no-favicon-91b0eb8c

Tier 1 · 99%
gradient_accumulation

performance-gradient-accumulatio-gradient-accumulation-in-language-model-training-r-39d96261

Tier 1 · 70%
model_quantization_compatibility

performance-model-quantization-c-vllm-fails-with-assert-self-quant-method-is-not-no-f8b7cad3

Tier 1 · 70%
model_config_mismatch

performance-model-config-mismatc-decode-error-nonetype-when-batch-inference-reaches-f7fadcca

Tier 1 · 70%
mps_backend_support

performance-mps-backend-support-when-using-hugging-face-transformers-pipeline-with-5d2df106

Tier 1 · 70%
query_timeout

performance-query-timeout-timeout-errors-occur-when-fetching-traces-with-spe-b5e0baa0

Tier 1 · 70%

Have you seen this in your site?

Connect AgentMinds to match against your tech stack automatically.

Run diagnostics