[NPU] NZ for non-quantized MOE, Qwen3 MOE double memory consumption fix (#15904)

Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>

docs/platforms/ascend_npu_deepseek_example.md

@@ -22,7 +22,6 @@ export SGLANG_ENABLE_OVERLAP_PLAN_STREAM=1
 #npu acceleration operator
 export SGLANG_NPU_USE_MLAPO=1
 export SGLANG_USE_FIA_NZ=1
-export ENABLE_MOE_NZ=1
 
 python3 -m sglang.launch_server \
     --model-path ${MODEL_PATH} \
@@ -71,7 +70,6 @@ export HCCL_BUFFSIZE=1536
 #npu acceleration operator
 export SGLANG_NPU_USE_MLAPO=1
 export SGLANG_USE_FIA_NZ=1
-export ENABLE_MOE_NZ=1
 export TASK_QUEUE_ENABLE=2
 
 python -m sglang.launch_server \

docs/platforms/ascend_npu_qwen3_examples.md

@@ -62,7 +62,6 @@ export HCCL_BUFFSIZE=1536
 export HCCL_OP_EXPANSION_MODE=AIV
 export SGLANG_DEEPEP_NUM_MAX_DISPATCH_TOKENS_PER_RANK=32
 export SGLANG_DEEPEP_BF16_DISPATCH=1
-export ENABLE_ASCEND_MOE_NZ=1
 
 python -m sglang.launch_server \
    --device npu \
@@ -84,7 +83,6 @@ export STREAMS_PER_DEVICE=32
 export HCCL_BUFFSIZE=1536
 export SGLANG_DEEPEP_NUM_MAX_DISPATCH_TOKENS_PER_RANK=32
 export SGLANG_DEEPEP_BF16_DISPATCH=1
-export ENABLE_ASCEND_MOE_NZ=1
 
 python -m sglang.launch_server \
    --model-path Qwen/Qwen3-235B-A22B-Instruct-2507 \