remove nccl pd mode.

docs/CN/source/tutorial/deepseek_deployment.rst

@@ -175,6 +175,7 @@ PD (Prefill-Decode) 分离模式将预填充和解码阶段分离部署，可以
 
     # PD prefill 模式 for DeepSeek-R1 (DP+EP) on H200
     # 使用方法: sh pd_prefill.sh <host> <pd_master_ip>
+    # 默认使用 NIXL 传输；如需使用 NCCL 数据面，可设置 LIGHTLLM_PD_KV_TRANSPORT_BACKEND=nccl
     # nvidia-cuda-mps-control -d，运行MPS(可选, 有mps支持性能会好特别多，但是部分显卡和驱动环境开启mps会容易出现错误，建议升级驱动到较高版本，特别是H系列卡)
 
     export host=$1
@@ -201,6 +202,7 @@ PD (Prefill-Decode) 分离模式将预填充和解码阶段分离部署，可以
 
     # PD decode 模式 for DeepSeek-R1 (DP+EP) on H200
     # 使用方法: sh pd_decode.sh <host> <pd_master_ip>
+    # 默认使用 NIXL 传输；如需使用 NCCL 数据面，可设置 LIGHTLLM_PD_KV_TRANSPORT_BACKEND=nccl
     export host=$1
     export pd_master_ip=$2
     nvidia-cuda-mps-control -d
@@ -336,4 +338,4 @@ PD (Prefill-Decode) 分离模式将预填充和解码阶段分离部署，可以
     --tokenizer_path /path/DeepSeek-R1/ \
     --url http://127.0.0.1:8088/generate_stream
 
-以上所有脚本可以参考 `test/start_scripts/multi_pd_master/` 目录下的脚本。
+以上所有脚本可以参考 `test/start_scripts/multi_pd_master/` 目录下的脚本。

docs/EN/source/tutorial/deepseek_deployment.rst

@@ -175,6 +175,7 @@ PD (Prefill-Decode) disaggregation mode separates prefill and decode stages for
 
     # PD prefill mode for DeepSeek-R1 (DP+EP) on H200
     # Usage: sh pd_prefill.sh <host> <pd_master_ip>
+    # NIXL is used by default. To use NCCL as the data-plane backend, set LIGHTLLM_PD_KV_TRANSPORT_BACKEND=nccl.
     # nvidia-cuda-mps-control -d, run MPS (optional, performance will be much better with mps support, but some GPUs may encounter errors when enabling mps, it's recommended to upgrade to a higher driver version, especially for H-series cards)
 
     export host=$1
@@ -198,6 +199,7 @@ PD (Prefill-Decode) disaggregation mode separates prefill and decode stages for
 
     # PD decode mode for DeepSeek-R1 (DP+EP) on H200
     # Usage: sh pd_decode.sh <host> <pd_master_ip>
+    # NIXL is used by default. To use NCCL as the data-plane backend, set LIGHTLLM_PD_KV_TRANSPORT_BACKEND=nccl.
     export host=$1
     export pd_master_ip=$2
     nvidia-cuda-mps-control -d
@@ -333,4 +335,4 @@ Supports multiple PD Master nodes, providing better load balancing and high avai
     --tokenizer_path /path/DeepSeek-R1/ \
     --url http://127.0.0.1:8088/generate_stream
 
-All the above scripts can be referenced from the scripts in the `test/start_scripts/multi_pd_master/` directory. 
+All the above scripts can be referenced from the scripts in the `test/start_scripts/multi_pd_master/` directory. 

lightllm/common/basemodel/basemodel.py

@@ -110,7 +110,6 @@ def __init__(self, kvargs):
         # 这可能会占用大量的显存，所以，req_manger 中保存的 mem_manger 是mem manager 初始化后再赋值
         self.req_manager.mem_manager = self.mem_manager
 
-        self._init_kv_move_buffer()
         self._check_mem_size()
         self._init_infer_layer()
         self._init_some_value()
@@ -197,11 +196,6 @@ def _init_mem_manager(self):
         )
         return
 
-    def _init_kv_move_buffer(self):
-        # p d 分离的推理模式下才需要做这一步初始化
-        if self.run_mode in ["prefill", "decode"]:
-            self.mem_manager.alloc_kv_move_buffer(self.mem_manager.size)
-
     def _check_mem_size(self):
         self.max_total_token_num = self.mem_manager.size