[3/N][Feat][Graph] Support all-to-all and quantized models with ACL Graph (#2614)

### What this PR does / why we need it?
* **Unify execution paths:** Consolidates the quantized and
non-quantized execution paths into a single `fused_experts` function,
removing duplicated logic and making the control flow clearer and easier
to maintain.
* **W8A8 dynamic quantization:** Adds support for W8A8 dynamic
quantization inside the unified MoE kernel. Communication routines are
updated to correctly handle dynamic quantization scales for activations.
* **Weight pre-processing:** Prae-transpose the `w13` and `w2` weight
matrices (as implemented in PR #2025) so that quantized and
non-quantized models follow the same code path for the MoE gating,
up-projection, and down-projection operations.
* **All-to-all communication:** Adds an `all-to-all` collective
communication pattern. For large token counts on modern hardware,
`all-to-all` is more efficient than the previous `all-gather` strategy.
However, `all-to-all` is not really captured and replayed due to
multiple D2H operations which will trigger synchronization, and thus
raise error when capture graphs. We only use `all-to-all` when fallback
to `compiled_graph_for_general_shape`.
* **Dynamic communication selection:** The model runner now selects the
optimal MoE communication method (`mc2`, `allgather`, or `alltoall`) at
runtime based on token count and the Ascend SoC version.
* **Limitation:** `all-gather` is not yet supported for quantized
models, which means there is still something left to do on A2.

### Does this PR introduce _any_ user-facing change?
None.

### How was this patch tested?
No further test cases needed.

- vLLM version: v0.10.1.1
- vLLM main:
d660c98c1b

---------

Signed-off-by: Yizhou Liu <liu_yizhou@outlook.com>

vllm_ascend/worker/model_runner_v1.py

@@ -89,7 +89,8 @@ from vllm_ascend.sample.rejection_sampler import AscendRejectionSampler
 from vllm_ascend.torchair.torchair_attention import AscendTorchairMetadata
 from vllm_ascend.torchair.torchair_mla import AscendMLATorchairMetadata
 from vllm_ascend.utils import (ACL_FORMAT_FRACTAL_ND, ACL_FORMAT_FRACTAL_NZ,
-                               ProfileExecuteDuration, is_310p,
+                               AscendSocVersion, ProfileExecuteDuration,
+                               get_ascend_soc_version, is_310p,
                                lmhead_tp_enable, vllm_version_is)
 from vllm_ascend.worker.eagle_proposer_v1 import EagleProposer
 from vllm_ascend.worker.mtp_proposer_v1 import MtpProposer
@@ -1620,8 +1621,22 @@ class NPUModelRunner(LoRAModelRunnerMixin):
         )
 
     def _select_moe_comm_method(self, num_tokens: int) -> str:
-        return ("mc2"
-                if num_tokens <= self.mc2_tokens_capacity else "allgather")
+        soc_version = get_ascend_soc_version()
+
+        if num_tokens <= self.mc2_tokens_capacity:
+            moe_comm_method = "mc2"
+        elif soc_version in {AscendSocVersion.A2}:
+            moe_comm_method = "allgather"
+        elif soc_version in {AscendSocVersion.A3}:
+            moe_comm_method = "alltoall"
+        else:
+            raise ValueError(f"Unsupported soc_version: {soc_version}")
+
+        if is_global_first_rank():
+            logger.debug(f"num_tokens: {num_tokens}, "
+                         f"moe_comm_method: {moe_comm_method}")
+
+        return moe_comm_method
 
     @torch.inference_mode()
     def execute_model(