EngineX/xc-llm-ascend
3
0
Fork 0
Code Issues Pull Requests Actions Projects Releases Wiki Activity

Revert PTA upgrade PR (#3352)

Browse Source 
we notice that torch npu 0919 doesn't work. This PR revert related
change which rely on 0919 version.
Revert PR: #3295  #3205  #3102 

Related: #3353

- vLLM version: v0.11.0
This commit is contained in:
wangxiyuan
2025-10-10 14:09:53 +08:00
committed by GitHub
parent 601a37aeff
commit ba19dd3183
15 changed files with 57 additions and 312 deletions
Download Patch File Download Diff File Expand all files Collapse all files

103
tests/ut/attention/test_attention_v1.py
View File

@@ -405,109 +405,6 @@ class TestAscendAttentionBackendImpl(TestBase):
mock_paged_attention.assert_called_once()
assert output.shape == (10, 8 * 64)
@patch('vllm_ascend.attention.attention_v1.get_forward_context')
@patch('vllm_ascend.attention.attention_v1.get_graph_params')
@patch('torch_npu._npu_reshape_and_cache')
@patch('torch_npu._npu_paged_attention')
@patch('torch.npu.graph_task_group_end')
@patch('torch.npu.graph_task_group_begin')
@patch('torch.npu.ExternalEvent')
@patch('torch_npu.npu.current_stream')
def test_paged_attention_with_existing_workspace(
self,
mock_get_forward_context,
mock_get_graph_params,
mock_npu_reshape_and_cache,
mock_paged_attention,
mock_graph_begin,
mock_graph_end,
mock_external_event_class,
mock_current_stream,
):
graph_params = MagicMock()
attn_metadata = MagicMock()
num_tokens = 10
graph_params.workspaces = {num_tokens: 10}
graph_params.events = {num_tokens: []}
graph_params.attn_params = {num_tokens: []}
graph_params.handles = {num_tokens: []}
query = torch.randn(2, 5, 8) # [batch_size, seq_len, hidden_size]
key_cache = MagicMock()
value_cache = MagicMock()
num_kv_heads = 4
num_heads = 8
scale = 0.1
output = torch.randn(2, 5, 8)
self_obj = MagicMock()
self_obj.key_cache = key_cache
self_obj.value_cache = value_cache
self_obj.num_kv_heads = num_kv_heads
self_obj.num_heads = num_heads
self_obj.scale = scale
mock_stream = MagicMock()
mock_current_stream.return_value = mock_stream
mock_event_instance = MagicMock()
mock_external_event_class.return_value = mock_event_instance
mock_handle = MagicMock()
mock_graph_end.return_value = mock_handle
workspace = graph_params.workspaces.get(num_tokens)
self.assertEqual(workspace, 10)
# 2. Handle graph capturing mode
stream = mock_current_stream()
event = mock_external_event_class()
event.wait(stream)
event.reset(stream)
graph_params.events[num_tokens].append(event)
graph_params.attn_params[num_tokens].append((
query,
self_obj.key_cache,
self_obj.value_cache,
self_obj.num_kv_heads,
self_obj.num_heads,
self_obj.scale,
attn_metadata.block_tables,
attn_metadata.seq_lens,
output,
))
mock_event_instance.wait.assert_called_once_with(mock_stream)
mock_event_instance.reset.assert_called_once_with(mock_stream)
self.assertEqual(len(graph_params.events[num_tokens]), 1)
self.assertEqual(len(graph_params.attn_params[num_tokens]), 1)
query = torch.randn(10, 8 * 64)
key = torch.randn(10, 8 * 64)
value = torch.randn(10, 8 * 64)
kv_cache = torch.empty(2, 5, 128, 8, 64)
metadata = self.attn_metadata
metadata.attn_state = AscendAttentionState.DecodeOnly
metadata.seq_lens = torch.tensor([10])
metadata.block_tables = torch.zeros(1, 5, dtype=torch.long)
metadata.num_actual_tokens = 10
metadata.slot_mapping = torch.zeros(10, dtype=torch.long)
layer = self.layer_no_quant
mock_get_forward_context.return_value = MagicMock(capturing=True)
mock_get_graph_params.return_value = graph_params
output = self.impl.forward(layer,
query,
key,
value,
kv_cache,
metadata,
trace_flag=False)
mock_paged_attention.assert_called_once()
self.assertEqual(len(graph_params.handles[num_tokens]), 0)
@patch('torch_npu._npu_reshape_and_cache')
@patch('torch_npu.npu_fused_infer_attention_score')
def test_forward_decode_only_swa(self, mock_fused_infer_attention_score,

24
tests/ut/ops/test_layernorm.py
View File

@@ -24,7 +24,7 @@ def mock_add_rms_norm(x, residual, weight, eps):
def mock_add_rms_norm_quant(x, residual, weight, quant_scale, quant_offset,
beta, epsilon):
epsilon):
x_out = 2 * x
residual_out = 2 * residual
x_out_quant = x_out.to(torch.int8)
@@ -94,7 +94,7 @@ class TestAscendRMSNorm(PytestBase):
mock_model_instance = mocker.MagicMock()
mock_forward_context.model_instance = mock_model_instance
mock_model_instance.model.layers = [
mocker.MagicMock() for _ in range(3)
mocker.MagicMock() for _ in range(2)
]
mock_layer_0 = mock_model_instance.model.layers[0]
@@ -124,7 +124,7 @@ class TestAscendRMSNorm(PytestBase):
mock_forward_context.addrmsnorm_quant_fusion_enabled = True
mock_forward_context.prefetch_mlp_enabled = False
mock_forward_context.layer_idx = 0
mock_forward_context.num_hidden_layers = 3
mock_forward_context.num_hidden_layers = 2
mock_forward_context.fusion_linear = "gate_up_dense"
# Ensure fusion and layer_idx increment are handled correctly
@@ -144,32 +144,18 @@ class TestAscendRMSNorm(PytestBase):
assert mock_forward_context.fusion_linear == "gate_up_dense"
assert mock_forward_context.layer_idx == 1
mock_forward_context.fusion_linear = "gate_moe"
x_out, residual_out = layer.forward_oot(x, residual)
assert mock_get_forward_context.call_count == 3
assert mock_forward_context.fusion_linear == "qkv_moe"
assert mock_forward_context.fusion_linear == "qkv_dense"
assert mock_forward_context.layer_idx == 2
x_out, residual_out = layer.forward_oot(x, residual)
assert mock_get_forward_context.call_count == 4
assert mock_forward_context.fusion_linear == "gate_moe"
assert mock_forward_context.fusion_linear == "qkv_dense"
assert mock_forward_context.layer_idx == 2
# last layer returned directly
x_out, residual_out = layer.forward_oot(x, residual)
assert mock_get_forward_context.call_count == 5
assert mock_forward_context.fusion_linear == "qkv_moe"
assert mock_forward_context.layer_idx == 3
x_out, residual_out = layer.forward_oot(x, residual)
assert mock_get_forward_context.call_count == 6
assert mock_forward_context.fusion_linear == "qkv_moe"
assert mock_forward_context.layer_idx == 3
if __name__ == '__main__':
unittest.main()

5
tests/ut/torchair/quantization/test_torchair_w8a8_dynamic.py
View File

@@ -1,6 +1,5 @@
from unittest.mock import MagicMock, patch
import pytest
import torch
from tests.ut.base import TestBase
@@ -17,10 +16,6 @@ class TestAscendW8A8FusedMoEMethod(TestBase):
self.hidden_size,
dtype=torch.bfloat16)
@pytest.mark.skipif(
True,
reason="fix me",
)
@patch("torch.distributed.all_to_all_single")
@patch("torch_npu.npu_moe_re_routing")
@patch("torch_npu.npu_grouped_matmul")