[2/N][Feat] Attention and MoE weight prefetch in Qwen3MoE models (#3203)

### What this PR does / why we need it?

- Refacotr and integrate a unified `WeightPrefetchMethod`
- Integrate `gate_up_proj.weight` in quantized Attention modules
- Prefetching these weights ahead of matmul-like operators imporves
performance by reducing L2 cache transfer latency

### Does this PR introduce _any_ user-facing change?

Add a new config in `--additional-config` for configuration:
```json
{
    "weight_prefetch_config": {
        "enabled": True,
        "prefetch_ratio": {
            "moe": {
                "gate_up": 0.8
            },
        },
    },
}
```
This feature is enabled by default, and can be disabled through this
configuration

### How was this patch tested?


- vLLM version: v0.11.0rc3
- vLLM main: https://github.com/vllm-project/vllm/commit/v0.11.0

---------

Signed-off-by: yuzhup <15705211260@163.com>

tests/ut/quantization/test_w8a8.py

@@ -755,6 +755,14 @@ class TestSelectExperts(TestBase):
         self.hidden_states = torch.randn(self.num_tokens, self.hidden_size)
         self.router_logits = torch.randn(self.num_tokens, self.num_experts)
 
+        self.mock_ctx = MagicMock()
+        self.mock_ctx.weight_prefetch_method = MagicMock()
+        patcher = patch(
+            'vllm_ascend.ops.moe.experts_selector.get_forward_context',
+            return_value=self.mock_ctx)
+        self.addCleanup(patcher.stop)
+        patcher.start()
+
     @patch('torch_npu.npu_moe_gating_top_k_softmax')
     def test_softmax_scoring(self, mock_topk):
         """Test softmax scoring function"""