add dispatch_gmm_combine kernel (#3532)

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

This PR introduces the Ascend implementation of the
`dispatch_ffn_combine` kernel and wires it into the vLLM-Ascend runtime,
together with follow‑up fixes to ensure the kernel builds and runs
correctly in CI.

- Add full host and device implementation of the `dispatch_ffn_combine`
kernel under `csrc/dispatch_ffn_combine`, including tiling logic, MOE
routing helpers, and kernel utilities for quantized FFN dispatch.
- Integrate the new kernel with the PyTorch binding
(csrc/torch_binding.cpp, csrc/torch_binding_meta.cpp) and the Ascend
runtime (vllm_ascend/ascend_forward_context.py,
vllm_ascend/worker/model_runner_v1.py).
- Extend fused MoE communication and token dispatch support in
`vllm_ascend/ops/fused_moe`, adding methods/utilities needed by the new
dispatch path.
- Update quantization logic in vllm_ascend/quantization/w8a8_dynamic.py
to support the new FFN dispatch flow.
- Fix kernel build issues by adjusting `csrc/build_aclnn.sh`, CMake
configuration, and include/namespace usage in the new kernel files.
- Add an end‑to‑end nightly test
`tests/e2e/nightly/ops/test_dispatch_ffn_combine.py` and helper
utilities in `vllm_ascend/utils.py` to validate the new kernel.

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

### How was this patch tested?


- vLLM version: v0.12.0
- vLLM main: https://github.com/vllm-project/vllm/commit/v0.12.0

---------

Signed-off-by: mojave2 <chenchen145@huawei.com>
Co-authored-by: wangxiyuan <wangxiyuan1007@gmail.com>

csrc/torch_binding_meta.cpp

@@ -156,7 +156,21 @@ void batch_matmul_transpose(const at::Tensor &tensor_a, const at::Tensor &tensor
                                     c10::optional<c10::string_view> quant_mode)
 {
     return;
+}
 
+at::Tensor& dispatch_ffn_combine_meta(
+    const at::Tensor& x,
+    const at::Tensor& weight1,
+    const at::Tensor& weight2,
+    const at::Tensor& expert_idx,
+    const at::Tensor& scale1,
+    const at::Tensor& scale2,
+    const at::Tensor& probs,
+    c10::string_view group,
+    int64_t max_output_size,
+    at::Tensor& out
+) {
+    return out;
 }
 
 at::Tensor npu_lightning_indexer_meta(
@@ -244,5 +258,7 @@ TORCH_LIBRARY_IMPL_EXPAND(CONCAT(_C, _ascend), Meta, ops) {
     ops.impl("npu_lightning_indexer", &vllm_ascend::meta::npu_lightning_indexer_meta);
     // Sparse flash attention
     ops.impl("npu_sparse_flash_attention", &vllm_ascend::meta::npu_sparse_flash_attention_meta);
+    // MoE dispatch-ffn-combine
+    ops.impl("dispatch_ffn_combine", &vllm_ascend::meta::dispatch_ffn_combine_meta);
 }
 }