The main purposes of this PR are as follows:
1. Remove the multicast-related code;
Reason:
1. In the scenario like a2 Dual-System Back-to-Back Networking,the
performance is worse than all_gather. Before the modification, in e2e
test, it was 3 tps; after the modification, it is 10 tps.
2. At the same time, we usually enable the SP feature,it is consistent
with the current logic.
3. The advantage of broadcast communication lies in the fact that it
does not suffer from uneven DP load and does not require the prefill ACL
graph to be enabled. But we support prefill Acl graph recently.
So we think there is no need to maintain the multicast as one choice in
moe communication.
Performance benefits are as follows:
When not enable_flashcomm1, TTFT remains relatively stable at around
43000ms, which is approximately 15000ms faster than before the
modification.
When enable_flashcomm1, there is no diffenence, TTFT remains relatively
stable at around 29000ms.
- vLLM version: v0.11.0
- vLLM main:
2918c1b49c
---------
Signed-off-by: weijinqian_v1 <weijinqian@huawei.com>
Signed-off-by: weijinqian0 <1184188277@qq.com>
Co-authored-by: weijinqian_v1 <weijinqian@huawei.com>
225 lines
8.6 KiB
Python
225 lines
8.6 KiB
Python
import unittest
|
|
from unittest.mock import MagicMock, patch
|
|
|
|
import torch
|
|
from vllm.model_executor.layers.fused_moe import FusedMoEConfig
|
|
|
|
from vllm_ascend.ops.fused_moe.prepare_finalize import (
|
|
PrepareAndFinalizeWithAll2All, PrepareAndFinalizeWithAllGather,
|
|
PrepareAndFinalizeWithMC2)
|
|
|
|
|
|
class TestPrepareAndFinalize(unittest.TestCase):
|
|
|
|
def setUp(self):
|
|
# Mock FusedMoEConfig
|
|
self.moe_config = MagicMock(spec=FusedMoEConfig)
|
|
self.moe_config.tp_group = MagicMock()
|
|
self.moe_config.tp_group.device_group = MagicMock()
|
|
self.moe_config.dp_size = 1
|
|
self.moe_config.tp_size = 1
|
|
self.moe_config.ep_size = 1
|
|
self.moe_config.dp_group = MagicMock()
|
|
self.moe_config.original_num_experts = 8
|
|
|
|
@patch(
|
|
"vllm_ascend.ops.fused_moe.prepare_finalize.get_tensor_model_parallel_world_size",
|
|
return_value=1)
|
|
@patch(
|
|
"vllm_ascend.ops.fused_moe.prepare_finalize.get_tensor_model_parallel_rank",
|
|
return_value=0)
|
|
@patch("vllm_ascend.ops.fused_moe.prepare_finalize.get_forward_context")
|
|
def test_mc2_prepare_finalize(self, mock_get_forward_context, mock_tp_rank,
|
|
mock_tp_size):
|
|
mock_context = MagicMock()
|
|
mock_context.mc2_mask = torch.tensor([1, 0, 1])
|
|
mock_context.padded_num_tokens = 4
|
|
mock_get_forward_context.return_value = mock_context
|
|
|
|
layer = PrepareAndFinalizeWithMC2(self.moe_config)
|
|
|
|
hidden_states = torch.randn(3, 8)
|
|
router_logits = torch.randn(3, 2)
|
|
|
|
h_out, r_out, mask, context_metadata = layer.prepare(
|
|
hidden_states, router_logits)
|
|
|
|
# Check padding and split
|
|
self.assertEqual(h_out.shape[0], 4)
|
|
self.assertEqual(r_out.shape[0], 4)
|
|
self.assertEqual(mask.tolist(), [1, 0, 1])
|
|
|
|
# Finalize
|
|
result = layer.finalize(h_out,
|
|
reduce_results=False,
|
|
context_metadata=context_metadata)
|
|
self.assertEqual(result.shape[0], 3)
|
|
|
|
@patch(
|
|
"vllm_ascend.ops.fused_moe.prepare_finalize.get_tensor_model_parallel_world_size",
|
|
return_value=2)
|
|
@patch(
|
|
"vllm_ascend.ops.fused_moe.prepare_finalize.get_tensor_model_parallel_rank",
|
|
return_value=0)
|
|
@patch("vllm_ascend.ops.fused_moe.prepare_finalize.get_forward_context")
|
|
@patch("torch.distributed.all_gather")
|
|
def test_mc2_tp_split_allgather(self, mock_all_gather,
|
|
mock_get_forward_context, mock_tp_rank,
|
|
mock_tp_size):
|
|
mock_context = MagicMock()
|
|
mock_context.mc2_mask = torch.tensor([1, 0, 1, 0])
|
|
mock_context.padded_num_tokens = 4
|
|
mock_get_forward_context.return_value = mock_context
|
|
|
|
layer = PrepareAndFinalizeWithMC2(self.moe_config)
|
|
hidden_states = torch.randn(4, 8)
|
|
router_logits = torch.randn(4, 2)
|
|
|
|
h_out, r_out, mask, context_metadata = layer.prepare(
|
|
hidden_states,
|
|
router_logits,
|
|
enable_shared_expert_dp=False,
|
|
replace_allreduce=False)
|
|
|
|
# With TP=2, should split into 2 parts
|
|
self.assertEqual(h_out.shape[0], 2)
|
|
|
|
# Mock all_gather behavior
|
|
def mock_all_gather_func(tensor_list, tensor, group=None):
|
|
tensor_list[0] = tensor
|
|
tensor_list[1] = tensor.clone()
|
|
|
|
mock_all_gather.side_effect = mock_all_gather_func
|
|
|
|
layer.split_hidden_states = [
|
|
torch.zeros_like(h_out),
|
|
torch.zeros_like(h_out)
|
|
]
|
|
final_result = layer.finalize(h_out,
|
|
reduce_results=False,
|
|
context_metadata=context_metadata)
|
|
|
|
# Should concat back to original size
|
|
self.assertEqual(final_result.shape[0], 4)
|
|
|
|
@patch(
|
|
"vllm_ascend.ops.fused_moe.prepare_finalize.get_tensor_model_parallel_world_size",
|
|
return_value=1)
|
|
@patch(
|
|
"vllm_ascend.ops.fused_moe.prepare_finalize.get_tensor_model_parallel_rank",
|
|
return_value=0)
|
|
def test_all2all_prepare_finalize(self, mock_tp_rank, mock_tp_size):
|
|
layer = PrepareAndFinalizeWithAll2All(self.moe_config)
|
|
hidden_states = torch.randn(3, 8)
|
|
router_logits = torch.randn(3, 2)
|
|
|
|
h_out, r_out, _, context_metadata = layer.prepare(
|
|
hidden_states, router_logits)
|
|
|
|
# Pad to tp_size=1, so no change
|
|
self.assertEqual(h_out.shape[0], 3)
|
|
|
|
result = layer.finalize(h_out,
|
|
reduce_results=False,
|
|
context_metadata=context_metadata)
|
|
self.assertEqual(result.shape[0], 3)
|
|
|
|
@patch(
|
|
"vllm_ascend.ops.fused_moe.prepare_finalize.get_tensor_model_parallel_world_size",
|
|
return_value=2)
|
|
@patch(
|
|
"vllm_ascend.ops.fused_moe.prepare_finalize.get_tensor_model_parallel_rank",
|
|
return_value=0)
|
|
@patch("torch.distributed.all_gather")
|
|
def test_all2all_tp_split_allgather(self, mock_all_gather, mock_tp_rank,
|
|
mock_tp_size):
|
|
layer = PrepareAndFinalizeWithAll2All(self.moe_config)
|
|
hidden_states = torch.randn(2, 8)
|
|
router_logits = torch.randn(2, 2)
|
|
|
|
h_out, r_out, _, context_metadata = layer.prepare(
|
|
hidden_states,
|
|
router_logits,
|
|
enable_shared_expert_dp=False,
|
|
replace_allreduce=False)
|
|
|
|
# Split due to TP=2
|
|
self.assertEqual(h_out.shape[0], 1)
|
|
|
|
# Mock all_gather
|
|
def mock_all_gather_func(tensor_list, tensor, group=None):
|
|
tensor_list[0] = tensor
|
|
tensor_list[1] = tensor.clone()
|
|
|
|
mock_all_gather.side_effect = mock_all_gather_func
|
|
|
|
layer.split_hidden_states = [
|
|
torch.zeros_like(h_out),
|
|
torch.zeros_like(h_out)
|
|
]
|
|
final_result = layer.finalize(h_out,
|
|
reduce_results=False,
|
|
context_metadata=context_metadata)
|
|
|
|
# Should concat back
|
|
self.assertEqual(final_result.shape[0], 2)
|
|
|
|
@patch("vllm_ascend.ops.fused_moe.prepare_finalize.get_dp_group")
|
|
@patch(
|
|
"vllm_ascend.ops.fused_moe.prepare_finalize.tensor_model_parallel_all_reduce"
|
|
)
|
|
@patch("vllm_ascend.ops.fused_moe.prepare_finalize.get_forward_context")
|
|
@patch("vllm_ascend.ops.fused_moe.prepare_finalize.enable_sp",
|
|
return_value=False)
|
|
def test_allgather_prepare_finalize(self, mock_enable_sp,
|
|
mock_get_forward_context,
|
|
mock_tp_all_reduce, mock_get_dp_group):
|
|
# Mock forward context
|
|
mock_context = MagicMock()
|
|
mock_context.max_tokens_across_dp = 6
|
|
mock_get_forward_context.return_value = mock_context
|
|
|
|
# Create a proper mock for DP group with working all_gather
|
|
mock_dp_group = MagicMock()
|
|
|
|
def mock_all_gather_func(tensor, dim):
|
|
# Simulate DP=2: repeat the tensor along the specified dimension
|
|
return torch.cat([tensor, tensor], dim=dim)
|
|
|
|
mock_dp_group.all_gather = mock_all_gather_func
|
|
mock_get_dp_group.return_value = mock_dp_group
|
|
|
|
self.moe_config.dp_size = 2
|
|
self.moe_config.tp_size = 1
|
|
self.moe_config.ep_size = 1
|
|
self.moe_config.dp_group = mock_dp_group
|
|
|
|
layer = PrepareAndFinalizeWithAllGather(self.moe_config)
|
|
|
|
hidden_states = torch.randn(3, 8)
|
|
router_logits = torch.randn(3, 2)
|
|
|
|
h_out, r_out, _, context_metadata = layer.prepare(
|
|
hidden_states, router_logits)
|
|
|
|
# After all-gather with DP=2, should double the batch size
|
|
self.assertEqual(h_out.shape[0], 12)
|
|
self.assertEqual(r_out.shape[0], 12)
|
|
|
|
# Finalize with reduce_scatter
|
|
def mock_reduce_scatter_func(tensor, dim):
|
|
# Simulate reduce_scatter: take first half
|
|
return tensor[:3]
|
|
|
|
mock_dp_group.reduce_scatter = mock_reduce_scatter_func
|
|
result = layer.finalize(h_out,
|
|
reduce_results=False,
|
|
context_metadata=context_metadata)
|
|
|
|
self.assertEqual(result.shape[0], 3)
|
|
|
|
# Test with TP all-reduce
|
|
mock_tp_all_reduce.return_value = result
|
|
result_with_tp = layer.finalize(h_out, reduce_results=True)
|
|
self.assertEqual(result_with_tp.shape[0], 3)
|