Revert "[feat] enable hierarchical mc2 ops on A2 by default (#5300)" (#5434)

We'll release 0.13.0 soon. The main branch is freeze. Let's revert the
newest change and redo it once 0.13.0 is released.

- vLLM version: release/v0.13.0
- vLLM main:
81786c8774

Signed-off-by: realliujiaxu <realliujiaxu@163.com>

vllm_ascend/ops/fused_moe/token_dispatcher.py

@@ -31,7 +31,8 @@ from vllm.distributed.parallel_state import get_ep_group
 from vllm_ascend.distributed.parallel_state import get_mc2_group
 from vllm_ascend.ops.fused_moe.comm_utils import (
     async_all_to_all, gather_from_sequence_parallel_region)
-from vllm_ascend.utils import AscendDeviceType, get_ascend_device_type
+from vllm_ascend.utils import (AscendDeviceType, get_ascend_device_type,
+                               is_hierarchical_communication_enabled)
 
 
 class MoETokenDispatcher(ABC):
@@ -100,6 +101,10 @@ class TokenDispatcherWithMC2(MoETokenDispatcher):
         self.need_extra_args = (
             get_ascend_device_type() == AscendDeviceType.A3)
 
+        # NOTE: When in A2, setting the environment variables HCCL_INTRA_PCIE_ENABLE=1 and
+        # HCCL_INTRA_ROCE_ENABLE=0 can reduce cross-machine communication traffic and significantly
+        # improve communication performance.
+        self.need_expert_scale = is_hierarchical_communication_enabled()
         self.with_quant = False
 
         # Here we need to calculate the global_bs = max_bs_per_rank * ep_world_size to execute
@@ -137,7 +142,6 @@ class TokenDispatcherWithMC2(MoETokenDispatcher):
         else:
             quant_mode = 0
             moe_expert_num = len(expert_map)
-
         kwargs_mc2 = {
             "x": hidden_states,
             "expert_ids": topk_ids,
@@ -146,12 +150,8 @@ class TokenDispatcherWithMC2(MoETokenDispatcher):
             "moe_expert_num": moe_expert_num,
             "global_bs": self.global_bs,
             "expert_token_nums_type": 0,
-            "expert_scales": topk_weights.to(torch.float32),
         }
 
-        if get_ascend_device_type() == AscendDeviceType.A2:
-            kwargs_mc2["comm_alg"] = "hierarchy"
-
         stage1_kwargs = {
             "scales": None,
             "quant_mode": quant_mode,
@@ -165,6 +165,11 @@ class TokenDispatcherWithMC2(MoETokenDispatcher):
                 "tp_world_size": 1,
                 "tp_rank_id": 0,
             })
+        if self.need_expert_scale:
+            stage1_kwargs.update({
+                "expert_scales":
+                topk_weights.to(torch.float32),
+            })
 
         kwargs_mc2.update(stage1_kwargs)
         return kwargs_mc2
@@ -264,12 +269,8 @@ class TokenDispatcherWithMC2(MoETokenDispatcher):
             "shared_expert_rank_num": 0,
             "moe_expert_num": moe_expert_num,
             "global_bs": self.global_bs,
-            "expand_scales": expand_scales,
         }
 
-        if get_ascend_device_type() == AscendDeviceType.A2:
-            kwargs_mc2["comm_alg"] = "hierarchy"
-
         if self.with_quant:
             tp_recv_counts = torch.empty(1,
                                          dtype=torch.int32,
@@ -280,6 +281,7 @@ class TokenDispatcherWithMC2(MoETokenDispatcher):
             "group_ep": self.moe_all_to_all_group_name,
             "ep_world_size": self.ep_world_size,
             "ep_rank_id": self.ep_rank_id,
+            "expand_scales": expand_scales,
         }
 
         if self.enable_dispatch_v2:

vllm_ascend/utils.py

@@ -958,6 +958,14 @@ def calculate_dp_buffer_size() -> int:
     return max(dp_buffer_size, _MIN_DP_BUFFER_SIZE)
 
 
+# Currently, when in A2, setting the environment variables HCCL_INTRA_PCIE_ENABLE=1
+# and HCCL_INTRA_ROCE_ENABLE=0 can reduce cross-machine communication traffic and
+# significantly improve communication performance of MC2 ops dispatch/combine.
+def is_hierarchical_communication_enabled():
+    return (os.getenv("HCCL_INTRA_ROCE_ENABLE", "") == "0"
+            and os.getenv("HCCL_INTRA_PCIE_ENABLE", "") == "1")
+
+
 def has_layer_idx(model_instance: torch.nn.Module) -> bool:
     if model_instance is None:
         return False