[Bug] InternVL3_5-8B stuck during starting up with tp=2

### Checklist

- [x] 1. I have searched related issues but cannot get the expected help.
- [x] 2. The bug has not been fixed in the latest version.
- [x] 3. Please note that if the bug-related issue you submitted lacks corresponding environment info and a minimal reproducible demo, it will be challenging for us to reproduce and resolve the issue, reducing the likelihood of receiving feedback.

### Describe the bug

InternVL3_5-8B runs well with tp=1 but stuck during starting up with tp=2. I found GPU-Util kept 100%, and 2 ray::IDLE process: %CPU 100%. I waited more than 1 hour, no changes.

workaround is : export NCCL_P2P_DISABLE=1

### Reproduction

```
lmdeploy serve api_server \
--server-port 8080 \
--tp 2 --backend pytorch \
--session-len 32768 \
/modles/InternVL3_5-8B
```

### Environment

```Shell
sys.platform: linux
Python: 3.10.18 (main, Jun 5 2025, 13:14:17) [GCC 11.2.0]
CUDA available: True
MUSA available: False
numpy_random_seed: 2147483648
GPU 0,1,2,3,4,5: NVIDIA RTX A6000
CUDA_HOME: /usr/local/cuda-12.6
NVCC: Cuda compilation tools, release 12.6, V12.6.85
GCC: gcc (Ubuntu 11.4.0-1ubuntu1~22.04) 11.4.0
PyTorch: 2.8.0+cu128
PyTorch compiling details: PyTorch built with:
- GCC 13.3
- C++ Version: 201703
- Intel(R) oneAPI Math Kernel Library Version 2024.2-Product Build 20240605 for Intel(R) 64 architecture applications
- Intel(R) MKL-DNN v3.7.1 (Git Hash 8d263e693366ef8db40acc569cc7d8edf644556d)
- OpenMP 201511 (a.k.a. OpenMP 4.5)
- LAPACK is enabled (usually provided by MKL)
- NNPACK is enabled
- CPU capability usage: AVX512
- CUDA Runtime 12.8
- NVCC architecture flags: -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;-gencode;arch=compute_100,code=sm_100;-gencode;arch=compute_120,code=sm_120
- CuDNN 91.0.2 (built against CUDA 12.9)
- Built with CuDNN 90.8
- Magma 2.6.1
- Build settings: BLAS_INFO=mkl, BUILD_TYPE=Release, COMMIT_SHA=a1cb3cc05d46d198467bebbb6e8fba50a325d4e7, CUDA_VERSION=12.8, CUDNN_VERSION=9.8.0, CXX_COMPILER=/opt/rh/gcc-toolset-13/root/usr/bin/c++, CXX_FLAGS= -fvisibility-inlines-hidden -DUSE_PTHREADPOOL -DNDEBUG -DUSE_KINETO -DLIBKINETO_NOROCTRACER -DLIBKINETO_NOXPUPTI=ON -DUSE_FBGEMM -DUSE_PYTORCH_QNNPACK -DUSE_XNNPACK -DSYMBOLICATE_MOBILE_DEBUG_HANDLE -O2 -fPIC -DC10_NODEPRECATED -Wall -Wextra -Werror=return-type -Werror=non-virtual-dtor -Werror=range-loop-construct -Werror=bool-operation -Wnarrowing -Wno-missing-field-initializers -Wno-unknown-pragmas -Wno-unused-parameter -Wno-strict-overflow -Wno-strict-aliasing -Wno-stringop-overflow -Wsuggest-override -Wno-psabi -Wno-error=old-style-cast -faligned-new -Wno-maybe-uninitialized -fno-math-errno -fno-trapping-math -Werror=format -Wno-dangling-reference -Wno-error=dangling-reference -Wno-stringop-overflow, LAPACK_INFO=mkl, PERF_WITH_AVX=1, PERF_WITH_AVX2=1, TORCH_VERSION=2.8.0, USE_CUDA=ON, USE_CUDNN=ON, USE_CUSPARSELT=1, USE_GFLAGS=OFF, USE_GLOG=OFF, USE_GLOO=ON, 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, USE_XCCL=OFF, USE_XPU=OFF,

TorchVision: 0.23.0+cu128
LMDeploy: 0.10.0+
transformers: 4.55.0
fastapi: 0.116.2
pydantic: 2.11.9
triton: 3.4.0
NVIDIA Topology:
GPU0 GPU1 GPU2 GPU3 GPU4 GPU5 CPU Affinity NUMA Affinity GPU NUMA ID
GPU0 X PHB PHB PHB PHB PHB 0-99 0-1 N/A
GPU1 PHB X PHB PHB PHB PHB 0-99 0-1 N/A
GPU2 PHB PHB X PHB PHB PHB 0-99 0-1 N/A
GPU3 PHB PHB PHB X PHB PHB 0-99 0-1 N/A
GPU4 PHB PHB PHB PHB X PHB 0-99 0-1 N/A
GPU5 PHB PHB PHB PHB PHB X 0-99 0-1 N/A

Legend:

X = Self
SYS = Connection traversing PCIe as well as the SMP interconnect between NUMA nodes (e.g., QPI/UPI)
NODE = Connection traversing PCIe as well as the interconnect between PCIe Host Bridges within a NUMA node
PHB = Connection traversing PCIe as well as a PCIe Host Bridge (typically the CPU)
PXB = Connection traversing multiple PCIe bridges (without traversing the PCIe Host Bridge)
PIX = Connection traversing at most a single PCIe bridge
NV# = Connection traversing a bonded set of # NVLinks
```

### Error traceback

```Shell
/app/Anaconda3/envs/lmdeploy/lib/python3.10/site-packages/torch/cuda/__init__.py:63: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
import pynvml # type: ignore[import]
2025-09-19 09:02:42,333 - lmdeploy - WARNING - __init__.py:10 - Disable DLSlime Backend
/app/Anaconda3/envs/lmdeploy/lib/python3.10/site-packages/torch/cuda/__init__.py:63: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
import pynvml # type: ignore[import]
2025-09-19 09:02:44,495 - lmdeploy - WARNING - __init__.py:10 - Disable DLSlime Backend
/app/Anaconda3/envs/lmdeploy/lib/python3.10/site-packages/torch/cuda/__init__.py:63: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
import pynvml # type: ignore[import]
2025-09-19 09:02:46,748 - lmdeploy - WARNING - __init__.py:10 - Disable DLSlime Backend
2025-09-19 09:02:47,448 - lmdeploy - WARNING - triton.py:24 - PytorchEngine has not been tested on triton>3.3.1.
2025-09-19 09:02:47,449 - lmdeploy - WARNING - transformers.py:22 - LMDeploy requires transformers version: [4.33.0 ~ 4.53.3], but found version: 4.55.0
2025-09-19 09:02:50,937 INFO worker.py:1951 -- Started a local Ray instance.
(pid=405497) /app/Anaconda3/envs/lmdeploy/lib/python3.10/site-packages/torch/cuda/__init__.py:63: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
(pid=405497) import pynvml # type: ignore[import]
```