[BugFix]Disable dispatch_gmm_combine_decode operator when mtp drafter model uses non-w8a8 while main model uses w8a8, or drafter model is eagle series

vllm_ascend/ascend_forward_context.py

@@ -14,7 +14,8 @@
 from vllm_ascend.ascend_config import get_ascend_config
 from vllm_ascend.utils import (AscendDeviceType, enable_sp, flashcomm2_enable,
                                get_ascend_device_type, has_layer_idx,
-                               is_moe_model)
+                               is_moe_model,
+                               speculative_enable_dispatch_gmm_combine_decode)
 
 
 class MoECommType(Enum):
@@ -242,14 +243,17 @@ def select_moe_comm_method(num_tokens: int,
         dynamic_eplb = ascend_config.dynamic_eplb or ascend_config.expert_map_record_path
         # TODO: drop the EP-size guard when dispatch_ffn_combine supports larger EP sizes
         # TODO: drop dynamic_eplb guard when dispatch_gmm_combine_decode supports tensor list inputs
-        # TODO: add guard for dispatch_gmm_combine_decode when mtp uses float while moe uses w8a8
+        # TODO: drop speculative method guard when dispatch_gmm_combine_decode supports w16a16
         fused_mc2_enable = envs_ascend.VLLM_ASCEND_ENABLE_FUSED_MC2 and quant_type == "w8a8_dynamic" and (
             not dynamic_eplb)
         if num_tokens <= mc2_tokens_capacity:
             fused_decode_enable = fused_mc2_enable
             if envs_ascend.VLLM_ASCEND_ENABLE_FUSED_MC2 == 1:
                 fused_decode_enable = fused_mc2_enable and get_ep_group(
                 ).world_size <= 16 and (not is_draft_model)
+            elif envs_ascend.VLLM_ASCEND_ENABLE_FUSED_MC2 == 2:
+                fused_decode_enable = fused_mc2_enable and \
+                    speculative_enable_dispatch_gmm_combine_decode(vllm_config)
             moe_comm_type = MoECommType.FUSED_MC2 if fused_decode_enable else MoECommType.MC2
         else:
             fused_prefill_enable = fused_mc2_enable

vllm_ascend/utils.py

@@ -823,6 +823,23 @@ def is_moe_model(vllm_config: VllmConfig):
     return _IS_MOE_MODEL
 
 
+def speculative_enable_dispatch_gmm_combine_decode(
+        vllm_config: VllmConfig) -> bool:
+    if vllm_config.speculative_config is None:
+        return True
+    speculative_method = getattr(vllm_config.speculative_config, "method",
+                                 None)
+    if speculative_method in [None, "ngram", "suffix"]:
+        return True
+    if speculative_method in ["eagle", "eagle3"]:
+        return False
+    if speculative_method == "mtp":
+        mtp_quant_type = getattr(vllm_config.model_config.hf_config,
+                                 "mtp_quantize", None)
+        return mtp_quant_type == "w8a8_dynamic"
+    return False
+
+
 def _is_contain_expert(config: Any):
     if isinstance(config, dict):
         for k, v in config.items():