[Bug] CUDA runtime error: an illegal memory access was encountered when 8bit kv quant was enabled

[aabbccddwasd]

Checklist

• 1. I have searched related issues but cannot get the expected help.
• 2. The bug has not been fixed in the latest version.

Describe the bug

I used this code to start a openAI server, it loaded successfully, but when I tried to generate some texts, it first output some texts and than raised a error "an illegal memory access was encountered"

this is the output↓

I looked on other issues, and I tried to set use_context_fmha=0 and to cache-max-entry-count to 0.1,neither of them worked, the only way I can solve it is to set quant-policy to 0 than it can run successfully.

I really want to use 8bit kv, is there somebody who can help me?

Reproduction

#!/bin/bash
export CUDA_VISIBLE_DEVICES=1,2,3,4
source /usr/local/anaconda3/bin/activate /home/aabbccddwasd/.conda/envs/obj013-env-lmdeploy
lmdeploy serve api_server ../obj013-qwen/Qwen1.5-110B-Chat-AWQ-turbomind \
    --server-name 0.0.0.0 \
    --server-port 8000 \
    --tp 4 \
    --model-format awq \
    --quant-policy 8 \
    --cache-max-entry-count 0.55

Environment

sys.platform: linux
Python: 3.12.4 | packaged by conda-forge | (main, Jun 17 2024, 10:23:07) [GCC 12.3.0]
CUDA available: True
MUSA available: False
numpy_random_seed: 2147483648
GPU 0,1,2,3,4: NVIDIA GeForce RTX 2080 Ti
CUDA_HOME: /usr/local/cuda-12.1
NVCC: Cuda compilation tools, release 12.1, V12.1.66
GCC: gcc (Ubuntu 12.3.0-1ubuntu1~22.04) 12.3.0
PyTorch: 2.2.2+cu121
PyTorch compiling details: PyTorch built with:
  - GCC 9.3
  - C++ Version: 201703
  - Intel(R) oneAPI Math Kernel Library Version 2022.2-Product Build 20220804 for Intel(R) 64 architecture applications
  - Intel(R) MKL-DNN v3.3.2 (Git Hash 2dc95a2ad0841e29db8b22fbccaf3e5da7992b01)
  - OpenMP 201511 (a.k.a. OpenMP 4.5)
  - LAPACK is enabled (usually provided by MKL)
  - NNPACK is enabled
  - CPU capability usage: AVX2
  - CUDA Runtime 12.1
  - NVCC architecture flags: -gencode;arch=compute_50,code=sm_50;-gencode;arch=compute_60,code=sm_60;-gencode;arch=compute_70,code=sm_70;-gencode;arch=compute_75,code=sm_75;-gencode;arch=compute_80,code=sm_80;-gencode;arch=compute_86,code=sm_86;-gencode;arch=compute_90,code=sm_90
  - CuDNN 8.9.2
  - Magma 2.6.1
  - Build settings: BLAS_INFO=mkl, BUILD_TYPE=Release, CUDA_VERSION=12.1, CUDNN_VERSION=8.9.2, CXX_COMPILER=/opt/rh/devtoolset-9/root/usr/bin/c++, CXX_FLAGS= -D_GLIBCXX_USE_CXX11_ABI=0 -fabi-version=11 -fvisibility-inlines-hidden -DUSE_PTHREADPOOL -DNDEBUG -DUSE_KINETO -DLIBKINETO_NOROCTRACER -DUSE_FBGEMM -DUSE_QNNPACK -DUSE_PYTORCH_QNNPACK -DUSE_XNNPACK -DSYMBOLICATE_MOBILE_DEBUG_HANDLE -O2 -fPIC -Wall -Wextra -Werror=return-type -Werror=non-virtual-dtor -Werror=bool-operation -Wnarrowing -Wno-missing-field-initializers -Wno-type-limits -Wno-array-bounds -Wno-unknown-pragmas -Wno-unused-parameter -Wno-unused-function -Wno-unused-result -Wno-strict-overflow -Wno-strict-aliasing -Wno-stringop-overflow -Wsuggest-override -Wno-psabi -Wno-error=pedantic -Wno-error=old-style-cast -Wno-missing-braces -fdiagnostics-color=always -faligned-new -Wno-unused-but-set-variable -Wno-maybe-uninitialized -fno-math-errno -fno-trapping-math -Werror=format -Wno-stringop-overflow, LAPACK_INFO=mkl, PERF_WITH_AVX=1, PERF_WITH_AVX2=1, PERF_WITH_AVX512=1, TORCH_VERSION=2.2.2, USE_CUDA=ON, USE_CUDNN=ON, USE_EXCEPTION_PTR=1, USE_GFLAGS=OFF, USE_GLOG=OFF, USE_MKL=ON, USE_MKLDNN=ON, USE_MPI=OFF, USE_NCCL=1, USE_NNPACK=ON, USE_OPENMP=ON, USE_ROCM=OFF, USE_ROCM_KERNEL_ASSERT=OFF, 

TorchVision: 0.17.2+cu121
LMDeploy: 0.4.2+
transformers: 4.42.2
gradio: Not Found
fastapi: 0.111.0
pydantic: 2.7.4
triton: 2.2.0

Error traceback

terminate called after throwing an instance of 'std::runtime_error'
  what():  [TM][ERROR] CUDA runtime error: an illegal memory access was encountered /lmdeploy/src/turbomind/models/llama/LlamaBatch.h:134 

lmdeploy-v1.5-110b-awq.sh: line 10: 699775 Aborted