fix neuron performance issue (#13589)

vllm/worker/neuron_worker.py

@@ -76,7 +76,7 @@ class NeuronWorker(LoraNotSupportedWorkerBase, LocalOrDistributedWorkerBase):
         # Set the number of GPU blocks to be the same as the maximum number of
         # sequences that can be processed in a single batch. This is equivalent
         # to schedule without PagedAttention.
-        num_gpu_blocks = self.scheduler_config.max_num_seqs
+        num_gpu_blocks = self.scheduler_config.max_num_seqs + 1
 
         # Swap not yet supported with Neuron backend.
         num_cpu_blocks = 0
@@ -90,7 +90,7 @@ class NeuronWorker(LoraNotSupportedWorkerBase, LocalOrDistributedWorkerBase):
 
         # Different values are not tested.
         assert num_cpu_blocks == 0
-        assert num_gpu_blocks == self.scheduler_config.max_num_seqs
+        assert num_gpu_blocks == self.scheduler_config.max_num_seqs + 1
 
         self.cache_config.num_gpu_blocks = num_gpu_blocks
         self.cache_config.num_cpu_blocks = num_cpu_blocks