8378bc28b0
### What this PR does / why we need it?
PCP/DCP splits the kv-cache onto different cards. After introducing the
parameter cp-kv-cache-interleave-size, the first size tokens will be
cached at Card 0, and so on.
However, if there are too few tokens, some cards will not store the
key-value pairs, resulting in values of 0, corrupted values, and
precision issues. Currently, additional operations are introduced to
avoid this precision problem.
After we integrate FIA operator in mla_cp._forward_decode and CANN
updates to 8.5.0, we now can remove these additional operations.
### Does this PR introduce _any_ user-facing change?
### How was this patch tested?
passed all CI by CANN 8.5.0
- vLLM version: v0.13.0
- vLLM main:
2c24bc6996
Signed-off-by: dsxsteven <dsxsteven@sina.com>
Signed-off-by: dsxsteven <36877507+dsxsteven@users.noreply.github.com>
281 lines
9.2 KiB
Python
281 lines
9.2 KiB
Python
#
|
|
# Copyright (c) 2025 Huawei Technologies Co., Ltd. All Rights Reserved.
|
|
# Copyright 2023 The vLLM team.
|
|
#
|
|
# Licensed under the Apache License, Version 2.0 (the "License");
|
|
# you may not use this file except in compliance with the License.
|
|
# You may obtain a copy of the License at
|
|
#
|
|
# http://www.apache.org/licenses/LICENSE-2.0
|
|
#
|
|
# Unless required by applicable law or agreed to in writing, software
|
|
# distributed under the License is distributed on an "AS IS" BASIS,
|
|
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
# See the License for the specific language governing permissions and
|
|
# limitations under the License.
|
|
#
|
|
"""
|
|
Compare the outputs of vLLM with and without context parallel.
|
|
|
|
Run `pytest tests/e2e/multicard/long_sequence/test_accuracy.py`.
|
|
"""
|
|
|
|
import pytest
|
|
|
|
from tests.e2e.conftest import VllmRunner
|
|
from tests.e2e.model_utils import check_outputs_equal
|
|
|
|
MODELS = [
|
|
"Qwen/Qwen3-8B",
|
|
"vllm-ascend/DeepSeek-V2-Lite-W8A8",
|
|
]
|
|
|
|
|
|
@pytest.mark.parametrize("model", MODELS)
|
|
@pytest.mark.parametrize("max_tokens", [10])
|
|
def test_models_long_sequence_output_between_tp_and_cp(
|
|
model: str,
|
|
max_tokens: int,
|
|
) -> None:
|
|
prompts = [
|
|
"The president of the United States is", "The capital of France is"
|
|
]
|
|
|
|
common_kwargs = {
|
|
"max_model_len": 1024,
|
|
}
|
|
|
|
if model == "vllm-ascend/DeepSeek-V2-Lite-W8A8":
|
|
cp_kwargs = {
|
|
"tensor_parallel_size": 2,
|
|
"decode_context_parallel_size": 2,
|
|
"prefill_context_parallel_size": 2,
|
|
"enable_expert_parallel": True,
|
|
"enforce_eager": True,
|
|
"quantization": "ascend",
|
|
}
|
|
tp_kwargs = {
|
|
"tensor_parallel_size": 4,
|
|
"enable_expert_parallel": True,
|
|
"enforce_eager": True,
|
|
"quantization": "ascend",
|
|
}
|
|
|
|
else:
|
|
cp_kwargs = {
|
|
"tensor_parallel_size": 1,
|
|
"decode_context_parallel_size": 1,
|
|
"prefill_context_parallel_size": 2,
|
|
"compilation_config": {
|
|
"cudagraph_mode": "FULL_DECODE_ONLY",
|
|
"cudagraph_capture_sizes": [4, 8, 24, 48, 60]
|
|
},
|
|
}
|
|
tp_kwargs = {
|
|
"tensor_parallel_size": 2,
|
|
"enforce_eager": True,
|
|
}
|
|
|
|
cp_full_kwargs = {}
|
|
cp_full_kwargs.update(common_kwargs) # type: ignore
|
|
cp_full_kwargs.update(cp_kwargs) # type: ignore
|
|
|
|
tp_full_kwargs = {}
|
|
tp_full_kwargs.update(common_kwargs) # type: ignore
|
|
tp_full_kwargs.update(tp_kwargs) # type: ignore
|
|
with VllmRunner(model, **cp_full_kwargs) as runner: # type: ignore
|
|
vllm_context_parallel_outputs = runner.generate_greedy(
|
|
prompts, max_tokens)
|
|
|
|
with VllmRunner(model, **tp_full_kwargs) as runner: # type: ignore
|
|
vllm_eager_outputs = runner.generate_greedy(prompts, max_tokens)
|
|
|
|
check_outputs_equal(
|
|
outputs_0_lst=vllm_eager_outputs,
|
|
outputs_1_lst=vllm_context_parallel_outputs,
|
|
name_0="vllm_eager_outputs",
|
|
name_1="vllm_context_parallel_outputs",
|
|
)
|
|
|
|
|
|
model = "vllm-ascend/DeepSeek-V2-Lite-W8A8"
|
|
|
|
|
|
@pytest.mark.parametrize("max_tokens", [10])
|
|
def test_accuracy_dcp_only_graph(max_tokens: int, ) -> None:
|
|
prompts = [
|
|
"The president of the United States is", "The capital of France is"
|
|
]
|
|
cp_kwargs = {
|
|
"tensor_parallel_size": 2,
|
|
"decode_context_parallel_size": 2,
|
|
"prefill_context_parallel_size": 1,
|
|
"enable_expert_parallel": True,
|
|
"compilation_config": {
|
|
"cudagraph_mode": "FULL_DECODE_ONLY",
|
|
"cudagraph_capture_sizes": [4, 8, 24, 48, 60]
|
|
},
|
|
"quantization": "ascend",
|
|
"max_model_len": 1024,
|
|
}
|
|
tp_kwargs = {
|
|
"tensor_parallel_size": 4,
|
|
"enable_expert_parallel": True,
|
|
"enforce_eager": True,
|
|
"quantization": "ascend",
|
|
"max_model_len": 1024,
|
|
}
|
|
with VllmRunner(model, **cp_kwargs) as runner: # type: ignore
|
|
vllm_context_parallel_outputs = runner.generate_greedy(
|
|
prompts, max_tokens)
|
|
|
|
with VllmRunner(model, **tp_kwargs) as runner: # type: ignore
|
|
vllm_eager_outputs = runner.generate_greedy(prompts, max_tokens)
|
|
|
|
check_outputs_equal(
|
|
outputs_0_lst=vllm_eager_outputs,
|
|
outputs_1_lst=vllm_context_parallel_outputs,
|
|
name_0="vllm_eager_outputs",
|
|
name_1="vllm_dcp_only_graph_outputs",
|
|
)
|
|
|
|
|
|
@pytest.mark.parametrize("max_tokens", [10])
|
|
def test_accuracy_dcp_only_eager(max_tokens: int, ) -> None:
|
|
prompts = [
|
|
"The president of the United States is", "The capital of France is"
|
|
]
|
|
cp_kwargs = {
|
|
"tensor_parallel_size": 2,
|
|
"decode_context_parallel_size": 2,
|
|
"prefill_context_parallel_size": 1,
|
|
"enable_expert_parallel": True,
|
|
"enforce_eager": True,
|
|
"quantization": "ascend",
|
|
"max_model_len": 1024,
|
|
}
|
|
tp_kwargs = {
|
|
"tensor_parallel_size": 4,
|
|
"enable_expert_parallel": True,
|
|
"enforce_eager": True,
|
|
"quantization": "ascend",
|
|
"max_model_len": 1024,
|
|
}
|
|
with VllmRunner(model, **cp_kwargs) as runner: # type: ignore
|
|
vllm_context_parallel_outputs = runner.generate_greedy(
|
|
prompts, max_tokens)
|
|
|
|
with VllmRunner(model, **tp_kwargs) as runner: # type: ignore
|
|
vllm_eager_outputs = runner.generate_greedy(prompts, max_tokens)
|
|
|
|
check_outputs_equal(
|
|
outputs_0_lst=vllm_eager_outputs,
|
|
outputs_1_lst=vllm_context_parallel_outputs,
|
|
name_0="vllm_eager_outputs",
|
|
name_1="vllm_dcp_only_eager_outputs",
|
|
)
|
|
|
|
|
|
@pytest.mark.parametrize("max_tokens", [10])
|
|
def test_accuracy_pcp_only(max_tokens: int, ) -> None:
|
|
prompts = [
|
|
"The president of the United States is", "The capital of France is"
|
|
]
|
|
cp_kwargs = {
|
|
"tensor_parallel_size": 2,
|
|
"decode_context_parallel_size": 1,
|
|
"prefill_context_parallel_size": 2,
|
|
"enable_expert_parallel": True,
|
|
"enforce_eager": True,
|
|
"quantization": "ascend",
|
|
"max_model_len": 1024,
|
|
}
|
|
tp_kwargs = {
|
|
"tensor_parallel_size": 4,
|
|
"enable_expert_parallel": True,
|
|
"enforce_eager": True,
|
|
"quantization": "ascend",
|
|
"max_model_len": 1024,
|
|
}
|
|
with VllmRunner(model, **cp_kwargs) as runner: # type: ignore
|
|
vllm_context_parallel_outputs = runner.generate_greedy(
|
|
prompts, max_tokens)
|
|
|
|
with VllmRunner(model, **tp_kwargs) as runner: # type: ignore
|
|
vllm_eager_outputs = runner.generate_greedy(prompts, max_tokens)
|
|
|
|
check_outputs_equal(
|
|
outputs_0_lst=vllm_eager_outputs,
|
|
outputs_1_lst=vllm_context_parallel_outputs,
|
|
name_0="vllm_eager_outputs",
|
|
name_1="vllm_pcp_only_outputs",
|
|
)
|
|
|
|
|
|
@pytest.mark.parametrize("model", MODELS)
|
|
@pytest.mark.parametrize("max_tokens", [10])
|
|
def test_models_long_sequence_cp_kv_interleave_size_output_between_tp_and_cp(
|
|
model: str,
|
|
max_tokens: int,
|
|
) -> None:
|
|
prompts = [
|
|
"The president of the United States is", "The capital of France is"
|
|
]
|
|
|
|
common_kwargs = {
|
|
"max_model_len": 1024,
|
|
}
|
|
|
|
if model == "vllm-ascend/DeepSeek-V2-Lite-W8A8":
|
|
cp_kwargs = {
|
|
"tensor_parallel_size": 2,
|
|
"decode_context_parallel_size": 2,
|
|
"prefill_context_parallel_size": 2,
|
|
"enable_expert_parallel": True,
|
|
"cp_kv_cache_interleave_size": 128,
|
|
"enforce_eager": True,
|
|
"quantization": "ascend",
|
|
}
|
|
tp_kwargs = {
|
|
"tensor_parallel_size": 4,
|
|
"enable_expert_parallel": True,
|
|
"enforce_eager": True,
|
|
"quantization": "ascend",
|
|
}
|
|
|
|
else:
|
|
cp_kwargs = {
|
|
"tensor_parallel_size": 1,
|
|
"decode_context_parallel_size": 1,
|
|
"prefill_context_parallel_size": 2,
|
|
"cp_kv_cache_interleave_size": 128,
|
|
"compilation_config": {
|
|
"cudagraph_mode": "FULL_DECODE_ONLY",
|
|
"cudagraph_capture_sizes": [4, 8, 24, 48, 60]
|
|
},
|
|
}
|
|
tp_kwargs = {
|
|
"tensor_parallel_size": 2,
|
|
"enforce_eager": True,
|
|
}
|
|
|
|
cp_full_kwargs = {}
|
|
cp_full_kwargs.update(common_kwargs) # type: ignore
|
|
cp_full_kwargs.update(cp_kwargs) # type: ignore
|
|
|
|
tp_full_kwargs = {}
|
|
tp_full_kwargs.update(common_kwargs) # type: ignore
|
|
tp_full_kwargs.update(tp_kwargs) # type: ignore
|
|
with VllmRunner(model, **cp_full_kwargs) as runner: # type: ignore
|
|
vllm_context_parallel_outputs = runner.generate_greedy(
|
|
prompts, max_tokens)
|
|
|
|
with VllmRunner(model, **tp_full_kwargs) as runner: # type: ignore
|
|
vllm_eager_outputs = runner.generate_greedy(prompts, max_tokens)
|
|
|
|
check_outputs_equal(
|
|
outputs_0_lst=vllm_eager_outputs,
|
|
outputs_1_lst=vllm_context_parallel_outputs,
|
|
name_0="vllm_eager_outputs",
|
|
name_1="vllm_context_parallel_outputs",
|
|
) |