2025-12-01 20:44:11 +08:00
|
|
|
from unittest.mock import patch
|
2025-08-14 17:18:30 +08:00
|
|
|
|
|
|
|
|
import pytest
|
|
|
|
|
import torch
|
|
|
|
|
from vllm.model_executor.layers.layernorm import RMSNorm
|
|
|
|
|
|
[refact] unified soc_version code (#4359)
### What this PR does / why we need it?
Currently, there are two paths to judge the chip type in code,
`get_ascend_soc_version` use `get_soc_version` api in torch_npu, and
`is_310p` `use _build_info.__soc_version__`, which generate when
install. We need to unify the two paths.
We need to unify these codes based on the following points:
1. We need to ensure consistency in chip type judgment between compiling
and running states;
2. In compiling state, we need chip type to complete op's compilation,
but in running state, we only need device
type(910B/910_93/310P/910_95/etc) to make code branch judgement;
3. In compiling state, torch_npu may not have been installed yet, so we
can't use torch_npu's api.
Based on the above points, we have made the following changes:
1. When user set env `SOC_VERSION`, use it; when not set, query
soc_version by `npu-smi`;
2. generate device_type based on soc_version when compiling, and write
`__device_type__` instead of `__soc_version__` in `_build_info.py`;
3. In running state, use `__device_type__` to judge code branch.
### Does this PR introduce _any_ user-facing change?
When not set env `SOC_VERSION`, it will not be `ASCEND910B1` by default,
we will query soc_version by `npu-smi`. And env `SOC_VERSION` must be in
the list `soc_to_device` in `setup.py`.
- vLLM version: v0.11.0
- vLLM main:
https://github.com/vllm-project/vllm/commit/2918c1b49c88c29783c86f78d2c4221cb9622379
Signed-off-by: zzzzwwjj <1183291235@qq.com>
2025-11-26 14:28:55 +08:00
|
|
|
from vllm_ascend.utils import AscendDeviceType
|
2025-08-14 17:18:30 +08:00
|
|
|
|
|
|
|
|
|
Adopt inductor fusion and define quantization fusion pass (#4168)
### What this PR does / why we need it?
The main goal of this PR to alleviate the high maintenance burden from
model duplication when we are going to do the model optimization. Some
of our optimized models diverges a little from the vllm's modeling, but
needs to rewrite several part of original one, brings negligible
maintenance bruden to the vllm-ascend.In order to solve that, we propose
to leverage `torch.compile` and `inductor pattern matcher`,
automatically fuse the pattern we want to merge. For more details can
refer to the RFC https://github.com/vllm-project/vllm-ascend/issues/4239
This pr integrates `AddRMSNorm` and the `Quant` operator, which can
improve the inference speed of models using `w8a8 `quantization.
### Does this PR introduce _any_ user-facing change?
Yes, add new additional_config
### How was this patch tested?
```python
def main():
prompts = [
"The president of the United States is Mr.",
]
# Create a sampling params object.
sampling_params = SamplingParams(max_tokens=100, temperature=0.6, top_k=40, top_p=0.95)
# Create an LLM.
llm = LLM(
model="/root/.cache/modelscope/hub/models/vllm-ascend/Qwen3-8B-W8A8",
# enforce_eager=True,
tensor_parallel_size=1,
trust_remote_code=True,
gpu_memory_utilization=0.7,
quantization="ascend",
)
# Generate texts from the prompts.
outputs = llm.generate(prompts, sampling_params)
for output in outputs:
prompt = output.prompt
generated_text = output.outputs[0].text
print(f"Prompt: {prompt!r}, Generated text: {generated_text!r}")
```
```text
Prompt: 'The president of the United States is Mr.', Generated text: ' Trump. The president of the United States is Mr. Biden. Which of the following statements is correct? \n\nA. Mr. Trump is Mr. Biden. \nB. Mr. Trump is not Mr. Biden. \nC. The president of the United States is not Mr. Trump. \nD. The president of the United States is not Mr. Biden.\n\nThe question presents a contradiction: it states that "The president of the United States is Mr. Trump" and "The president of'
```
- vLLM version: 86e178f7c4d8c3b0eaf3c8e3f810a83f63b90e24
- vLLM main:
https://github.com/vllm-project/vllm/commit/86e178f7c4d8c3b0eaf3c8e3f810a83f63b90e24
---------
Signed-off-by: Icey <1790571317@qq.com>
Signed-off-by: wxsIcey <1790571317@qq.com>
2025-12-04 10:29:48 +08:00
|
|
|
@pytest.fixture
|
|
|
|
|
def dummy_tensor():
|
|
|
|
|
return torch.randn(4, 8, dtype=torch.float16)
|
|
|
|
|
|
|
|
|
|
|
2025-08-14 17:18:30 +08:00
|
|
|
def mock_rms_norm(x, weight, eps):
|
|
|
|
|
return x + 1, None
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
def mock_add_rms_norm(x, residual, weight, eps):
|
|
|
|
|
return 2 * x, None, 2 * residual
|
|
|
|
|
|
|
|
|
|
|
Adopt inductor fusion and define quantization fusion pass (#4168)
### What this PR does / why we need it?
The main goal of this PR to alleviate the high maintenance burden from
model duplication when we are going to do the model optimization. Some
of our optimized models diverges a little from the vllm's modeling, but
needs to rewrite several part of original one, brings negligible
maintenance bruden to the vllm-ascend.In order to solve that, we propose
to leverage `torch.compile` and `inductor pattern matcher`,
automatically fuse the pattern we want to merge. For more details can
refer to the RFC https://github.com/vllm-project/vllm-ascend/issues/4239
This pr integrates `AddRMSNorm` and the `Quant` operator, which can
improve the inference speed of models using `w8a8 `quantization.
### Does this PR introduce _any_ user-facing change?
Yes, add new additional_config
### How was this patch tested?
```python
def main():
prompts = [
"The president of the United States is Mr.",
]
# Create a sampling params object.
sampling_params = SamplingParams(max_tokens=100, temperature=0.6, top_k=40, top_p=0.95)
# Create an LLM.
llm = LLM(
model="/root/.cache/modelscope/hub/models/vllm-ascend/Qwen3-8B-W8A8",
# enforce_eager=True,
tensor_parallel_size=1,
trust_remote_code=True,
gpu_memory_utilization=0.7,
quantization="ascend",
)
# Generate texts from the prompts.
outputs = llm.generate(prompts, sampling_params)
for output in outputs:
prompt = output.prompt
generated_text = output.outputs[0].text
print(f"Prompt: {prompt!r}, Generated text: {generated_text!r}")
```
```text
Prompt: 'The president of the United States is Mr.', Generated text: ' Trump. The president of the United States is Mr. Biden. Which of the following statements is correct? \n\nA. Mr. Trump is Mr. Biden. \nB. Mr. Trump is not Mr. Biden. \nC. The president of the United States is not Mr. Trump. \nD. The president of the United States is not Mr. Biden.\n\nThe question presents a contradiction: it states that "The president of the United States is Mr. Trump" and "The president of'
```
- vLLM version: 86e178f7c4d8c3b0eaf3c8e3f810a83f63b90e24
- vLLM main:
https://github.com/vllm-project/vllm/commit/86e178f7c4d8c3b0eaf3c8e3f810a83f63b90e24
---------
Signed-off-by: Icey <1790571317@qq.com>
Signed-off-by: wxsIcey <1790571317@qq.com>
2025-12-04 10:29:48 +08:00
|
|
|
@pytest.mark.parametrize("is_310p", [True, False])
|
|
|
|
|
@pytest.mark.parametrize("residual",
|
|
|
|
|
[None, torch.randn(4, 8, dtype=torch.float32)])
|
|
|
|
|
@patch("torch_npu.npu_rms_norm", side_effect=mock_rms_norm)
|
|
|
|
|
@patch("torch_npu.npu_add_rms_norm", side_effect=mock_add_rms_norm)
|
|
|
|
|
def test_RMSNorm_forward(mock_add_rmsnorm, mock_rmsnorm, is_310p, residual,
|
|
|
|
|
dummy_tensor):
|
2025-09-16 22:31:38 +08:00
|
|
|
|
Adopt inductor fusion and define quantization fusion pass (#4168)
### What this PR does / why we need it?
The main goal of this PR to alleviate the high maintenance burden from
model duplication when we are going to do the model optimization. Some
of our optimized models diverges a little from the vllm's modeling, but
needs to rewrite several part of original one, brings negligible
maintenance bruden to the vllm-ascend.In order to solve that, we propose
to leverage `torch.compile` and `inductor pattern matcher`,
automatically fuse the pattern we want to merge. For more details can
refer to the RFC https://github.com/vllm-project/vllm-ascend/issues/4239
This pr integrates `AddRMSNorm` and the `Quant` operator, which can
improve the inference speed of models using `w8a8 `quantization.
### Does this PR introduce _any_ user-facing change?
Yes, add new additional_config
### How was this patch tested?
```python
def main():
prompts = [
"The president of the United States is Mr.",
]
# Create a sampling params object.
sampling_params = SamplingParams(max_tokens=100, temperature=0.6, top_k=40, top_p=0.95)
# Create an LLM.
llm = LLM(
model="/root/.cache/modelscope/hub/models/vllm-ascend/Qwen3-8B-W8A8",
# enforce_eager=True,
tensor_parallel_size=1,
trust_remote_code=True,
gpu_memory_utilization=0.7,
quantization="ascend",
)
# Generate texts from the prompts.
outputs = llm.generate(prompts, sampling_params)
for output in outputs:
prompt = output.prompt
generated_text = output.outputs[0].text
print(f"Prompt: {prompt!r}, Generated text: {generated_text!r}")
```
```text
Prompt: 'The president of the United States is Mr.', Generated text: ' Trump. The president of the United States is Mr. Biden. Which of the following statements is correct? \n\nA. Mr. Trump is Mr. Biden. \nB. Mr. Trump is not Mr. Biden. \nC. The president of the United States is not Mr. Trump. \nD. The president of the United States is not Mr. Biden.\n\nThe question presents a contradiction: it states that "The president of the United States is Mr. Trump" and "The president of'
```
- vLLM version: 86e178f7c4d8c3b0eaf3c8e3f810a83f63b90e24
- vLLM main:
https://github.com/vllm-project/vllm/commit/86e178f7c4d8c3b0eaf3c8e3f810a83f63b90e24
---------
Signed-off-by: Icey <1790571317@qq.com>
Signed-off-by: wxsIcey <1790571317@qq.com>
2025-12-04 10:29:48 +08:00
|
|
|
with patch("vllm_ascend.utils.get_ascend_device_type",
|
|
|
|
|
return_value=AscendDeviceType._310P
|
|
|
|
|
if is_310p else AscendDeviceType._910_93):
|
2025-09-08 22:52:24 +08:00
|
|
|
layer = RMSNorm(hidden_size=8, eps=1e-05)
|
2025-08-14 17:18:30 +08:00
|
|
|
if residual is not None:
|
Adopt inductor fusion and define quantization fusion pass (#4168)
### What this PR does / why we need it?
The main goal of this PR to alleviate the high maintenance burden from
model duplication when we are going to do the model optimization. Some
of our optimized models diverges a little from the vllm's modeling, but
needs to rewrite several part of original one, brings negligible
maintenance bruden to the vllm-ascend.In order to solve that, we propose
to leverage `torch.compile` and `inductor pattern matcher`,
automatically fuse the pattern we want to merge. For more details can
refer to the RFC https://github.com/vllm-project/vllm-ascend/issues/4239
This pr integrates `AddRMSNorm` and the `Quant` operator, which can
improve the inference speed of models using `w8a8 `quantization.
### Does this PR introduce _any_ user-facing change?
Yes, add new additional_config
### How was this patch tested?
```python
def main():
prompts = [
"The president of the United States is Mr.",
]
# Create a sampling params object.
sampling_params = SamplingParams(max_tokens=100, temperature=0.6, top_k=40, top_p=0.95)
# Create an LLM.
llm = LLM(
model="/root/.cache/modelscope/hub/models/vllm-ascend/Qwen3-8B-W8A8",
# enforce_eager=True,
tensor_parallel_size=1,
trust_remote_code=True,
gpu_memory_utilization=0.7,
quantization="ascend",
)
# Generate texts from the prompts.
outputs = llm.generate(prompts, sampling_params)
for output in outputs:
prompt = output.prompt
generated_text = output.outputs[0].text
print(f"Prompt: {prompt!r}, Generated text: {generated_text!r}")
```
```text
Prompt: 'The president of the United States is Mr.', Generated text: ' Trump. The president of the United States is Mr. Biden. Which of the following statements is correct? \n\nA. Mr. Trump is Mr. Biden. \nB. Mr. Trump is not Mr. Biden. \nC. The president of the United States is not Mr. Trump. \nD. The president of the United States is not Mr. Biden.\n\nThe question presents a contradiction: it states that "The president of the United States is Mr. Trump" and "The president of'
```
- vLLM version: 86e178f7c4d8c3b0eaf3c8e3f810a83f63b90e24
- vLLM main:
https://github.com/vllm-project/vllm/commit/86e178f7c4d8c3b0eaf3c8e3f810a83f63b90e24
---------
Signed-off-by: Icey <1790571317@qq.com>
Signed-off-by: wxsIcey <1790571317@qq.com>
2025-12-04 10:29:48 +08:00
|
|
|
out_x, out_residual = layer.forward_oot(dummy_tensor, residual)
|
|
|
|
|
|
|
|
|
|
if is_310p:
|
|
|
|
|
expected_arg_x = dummy_tensor + residual.to(dummy_tensor.dtype)
|
|
|
|
|
expected_out_x = expected_arg_x + 1
|
|
|
|
|
expected_out_residual = expected_arg_x.to(residual.dtype)
|
|
|
|
|
|
|
|
|
|
mock_rmsnorm.assert_called_once()
|
|
|
|
|
assert torch.allclose(out_x, expected_out_x)
|
|
|
|
|
assert torch.allclose(out_residual, expected_out_residual)
|
|
|
|
|
else:
|
|
|
|
|
expected_out_x = 2 * dummy_tensor
|
|
|
|
|
expected_out_residual = 2 * residual
|
|
|
|
|
mock_add_rmsnorm.assert_called_once()
|
|
|
|
|
assert torch.allclose(out_x, expected_out_x)
|
|
|
|
|
assert torch.allclose(out_residual, expected_out_residual)
|
2025-08-14 17:18:30 +08:00
|
|
|
else:
|
Adopt inductor fusion and define quantization fusion pass (#4168)
### What this PR does / why we need it?
The main goal of this PR to alleviate the high maintenance burden from
model duplication when we are going to do the model optimization. Some
of our optimized models diverges a little from the vllm's modeling, but
needs to rewrite several part of original one, brings negligible
maintenance bruden to the vllm-ascend.In order to solve that, we propose
to leverage `torch.compile` and `inductor pattern matcher`,
automatically fuse the pattern we want to merge. For more details can
refer to the RFC https://github.com/vllm-project/vllm-ascend/issues/4239
This pr integrates `AddRMSNorm` and the `Quant` operator, which can
improve the inference speed of models using `w8a8 `quantization.
### Does this PR introduce _any_ user-facing change?
Yes, add new additional_config
### How was this patch tested?
```python
def main():
prompts = [
"The president of the United States is Mr.",
]
# Create a sampling params object.
sampling_params = SamplingParams(max_tokens=100, temperature=0.6, top_k=40, top_p=0.95)
# Create an LLM.
llm = LLM(
model="/root/.cache/modelscope/hub/models/vllm-ascend/Qwen3-8B-W8A8",
# enforce_eager=True,
tensor_parallel_size=1,
trust_remote_code=True,
gpu_memory_utilization=0.7,
quantization="ascend",
)
# Generate texts from the prompts.
outputs = llm.generate(prompts, sampling_params)
for output in outputs:
prompt = output.prompt
generated_text = output.outputs[0].text
print(f"Prompt: {prompt!r}, Generated text: {generated_text!r}")
```
```text
Prompt: 'The president of the United States is Mr.', Generated text: ' Trump. The president of the United States is Mr. Biden. Which of the following statements is correct? \n\nA. Mr. Trump is Mr. Biden. \nB. Mr. Trump is not Mr. Biden. \nC. The president of the United States is not Mr. Trump. \nD. The president of the United States is not Mr. Biden.\n\nThe question presents a contradiction: it states that "The president of the United States is Mr. Trump" and "The president of'
```
- vLLM version: 86e178f7c4d8c3b0eaf3c8e3f810a83f63b90e24
- vLLM main:
https://github.com/vllm-project/vllm/commit/86e178f7c4d8c3b0eaf3c8e3f810a83f63b90e24
---------
Signed-off-by: Icey <1790571317@qq.com>
Signed-off-by: wxsIcey <1790571317@qq.com>
2025-12-04 10:29:48 +08:00
|
|
|
out_x = layer.forward_oot(dummy_tensor, residual)
|
|
|
|
|
expected_out_x = dummy_tensor + 1
|
2025-09-16 22:31:38 +08:00
|
|
|
|
Adopt inductor fusion and define quantization fusion pass (#4168)
### What this PR does / why we need it?
The main goal of this PR to alleviate the high maintenance burden from
model duplication when we are going to do the model optimization. Some
of our optimized models diverges a little from the vllm's modeling, but
needs to rewrite several part of original one, brings negligible
maintenance bruden to the vllm-ascend.In order to solve that, we propose
to leverage `torch.compile` and `inductor pattern matcher`,
automatically fuse the pattern we want to merge. For more details can
refer to the RFC https://github.com/vllm-project/vllm-ascend/issues/4239
This pr integrates `AddRMSNorm` and the `Quant` operator, which can
improve the inference speed of models using `w8a8 `quantization.
### Does this PR introduce _any_ user-facing change?
Yes, add new additional_config
### How was this patch tested?
```python
def main():
prompts = [
"The president of the United States is Mr.",
]
# Create a sampling params object.
sampling_params = SamplingParams(max_tokens=100, temperature=0.6, top_k=40, top_p=0.95)
# Create an LLM.
llm = LLM(
model="/root/.cache/modelscope/hub/models/vllm-ascend/Qwen3-8B-W8A8",
# enforce_eager=True,
tensor_parallel_size=1,
trust_remote_code=True,
gpu_memory_utilization=0.7,
quantization="ascend",
)
# Generate texts from the prompts.
outputs = llm.generate(prompts, sampling_params)
for output in outputs:
prompt = output.prompt
generated_text = output.outputs[0].text
print(f"Prompt: {prompt!r}, Generated text: {generated_text!r}")
```
```text
Prompt: 'The president of the United States is Mr.', Generated text: ' Trump. The president of the United States is Mr. Biden. Which of the following statements is correct? \n\nA. Mr. Trump is Mr. Biden. \nB. Mr. Trump is not Mr. Biden. \nC. The president of the United States is not Mr. Trump. \nD. The president of the United States is not Mr. Biden.\n\nThe question presents a contradiction: it states that "The president of the United States is Mr. Trump" and "The president of'
```
- vLLM version: 86e178f7c4d8c3b0eaf3c8e3f810a83f63b90e24
- vLLM main:
https://github.com/vllm-project/vllm/commit/86e178f7c4d8c3b0eaf3c8e3f810a83f63b90e24
---------
Signed-off-by: Icey <1790571317@qq.com>
Signed-off-by: wxsIcey <1790571317@qq.com>
2025-12-04 10:29:48 +08:00
|
|
|
mock_rmsnorm.assert_called_once()
|
|
|
|
|
assert torch.allclose(out_x, expected_out_x)
|