From 7a4a5de729466de4142e70b845e78a0171296468 Mon Sep 17 00:00:00 2001
From: Chauncey <chaunceyjiang@gmail.com>
Date: Fri, 18 Apr 2025 13:12:42 +0800
Subject: [PATCH] [Misc] Update outdated note: LMCache now supports chunked
 prefill (#16697)

Signed-off-by: chaunceyjiang <chaunceyjiang@gmail.com>
---
 examples/offline_inference/cpu_offload_lmcache.py | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/examples/offline_inference/cpu_offload_lmcache.py b/examples/offline_inference/cpu_offload_lmcache.py
index 025444233d3b0..37aea281032fd 100644
--- a/examples/offline_inference/cpu_offload_lmcache.py
+++ b/examples/offline_inference/cpu_offload_lmcache.py
@@ -37,11 +37,11 @@ def build_llm_with_lmcache():
         '{"kv_connector":"LMCacheConnector", "kv_role":"kv_both"}')
     # Set GPU memory utilization to 0.8 for an A40 GPU with 40GB
     # memory. Reduce the value if your GPU has less memory.
-    # Note that LMCache is not compatible with chunked prefill for now.
+    # Note: LMCache supports chunked prefill (see vLLM#14505, LMCache#392).
     llm = LLM(model="mistralai/Mistral-7B-Instruct-v0.2",
               kv_transfer_config=ktc,
               max_model_len=8000,
-              enable_chunked_prefill=False,
+              enable_chunked_prefill=True,
               gpu_memory_utilization=0.8)
 
     try: