Speed up expert location update (#6661)

python/sglang/srt/model_executor/expert_location_updater.py

@@ -12,7 +12,6 @@
 # limitations under the License.
 # ==============================================================================
 import logging
-from datetime import timedelta
 from typing import Dict, List, Tuple
 
 import torch
@@ -348,14 +347,8 @@ def update_expert_weights_single_layer(
             return
 
         reqs = torch.distributed.batch_isend_irecv(p2p_ops)
-        try:
         for req in reqs:
-                req.wait(timeout=timedelta(seconds=30))
-        except RuntimeError:
-            logger.error(
-                f"Context: {rank=} {old_physical_to_logical_map=} {new_physical_to_logical_map=} {num_local_physical_experts=} {num_gpu_per_node=}"
-            )
-            raise
+            req.wait()
 
     def _execute_buffer2weight_copies(buffer2weight_copy_infos):
         for (