9fdabb7b60
This PR introduces the `EXEC_NPU_CMD` macro, serving as an adapter layer to simplify the invocation of `aclnn` operators on Ascend NPUs. **Key Changes:** * **Adapter Layer:** Added `EXEC_NPU_CMD` macro and related dependencies to standardize `aclnn` calls. * **Operator Support:** Integrated `grouped_matmul_swiglu_quant` as a reference implementation to demonstrate the usage of the new macro. --- - vLLM version: v0.11.2 --------- Signed-off-by: SlightwindSec <slightwindsec@gmail.com>
31 lines
942 B
C++
31 lines
942 B
C++
// Copyright (c) 2020, Huawei Technologies Co., Ltd
|
|
// All rights reserved.
|
|
//
|
|
// This source code is licensed under the BSD-style license found in the
|
|
// LICENSE file in the root directory of this source tree.
|
|
|
|
#include "NPUBridge.h"
|
|
|
|
namespace vllm_ascend
|
|
{
|
|
NPUStorageImpl *NPUBridge::GetNpuStorageImpl(c10::StorageImpl *storageImpl)
|
|
{
|
|
return static_cast<NPUStorageImpl *>(storageImpl);
|
|
}
|
|
|
|
NPUStorageImpl *NPUBridge::GetNpuStorageImpl(c10::Storage &&storage)
|
|
{
|
|
return static_cast<NPUStorageImpl *>(storage.unsafeGetStorageImpl());
|
|
}
|
|
|
|
NPUStorageImpl *NPUBridge::GetNpuStorageImpl(const at::Tensor &tensor)
|
|
{
|
|
return static_cast<NPUStorageImpl *>(tensor.storage().unsafeGetStorageImpl());
|
|
}
|
|
|
|
NPUStorageDesc &NPUBridge::GetNpuStorageImplDesc(const at::Tensor &tensor)
|
|
{
|
|
return static_cast<NPUStorageImpl *>(tensor.storage().unsafeGetStorageImpl())->npu_desc_;
|
|
}
|
|
}
|