wehub-resource-sync
239 lines
8.1 KiB
Markdown
239 lines
8.1 KiB
Markdown
# 自定义后端
|
||
Runtime-Backend是MNN对计算设备的抽象。MNN当前已经支持CPU、Vulkan、OpenCL、Metal、CUDA等Backend，**只在计算设备暂未支持时新增Backend**，新增Op，请参阅[新增Op文档](op)。
|
||
|
||
|
||
|
||
## 声明
|
||
所有新增Backend都需继承`Backend`类，并实现所有纯虚函数。
|
||
```cpp
|
||
class XPUBackend final : public Backend {
|
||
XPUBackend(MNNForwardType type, MemoryMode mode);
|
||
virtual ~XPUBackend();
|
||
virtual Execution* onCreate(const std::vector<Tensor*>& inputs, const std::vector<Tensor*>& outputs, const MNN::Op* op) override;
|
||
virtual void onExecuteBegin() const override;
|
||
virtual void onExecuteEnd() const override;
|
||
virtual void onResizeBegin() override;
|
||
virtual ErrorCode onResizeEnd() override;
|
||
|
||
virtual MemObj* onAcquire(const Tensor* tensor, StorageType storageType) override;
|
||
virtual bool onClearBuffer() override;
|
||
virtual void onCopyBuffer(const Tensor* srcTensor, const Tensor* dstTensor) const override;
|
||
}
|
||
```
|
||
|
||
## 构造与销毁
|
||
Backend构造时，可以额外指定内存环境，在内存受限环境中，应避免非必要的内存使用。可以在构造函数中，完成对计算设备访问的必要初始化，如GPU下预加载shader等。
|
||
```cpp
|
||
/** backend memory mode */
|
||
enum MemoryMode {
|
||
/** use memory without limit. */
|
||
NORMAL = 0,
|
||
/** use memory thriftily. */
|
||
LIMIT = 1
|
||
};
|
||
/**
|
||
* @brief initializer.
|
||
* @param type forward type.
|
||
* @param mode memory mode.
|
||
*/
|
||
Backend(MNNForwardType type, MemoryMode mode = NORMAL);
|
||
```
|
||
|
||
## Execution创建
|
||
Backend需要通过`onCreate`为op创建出exection实例：
|
||
```cpp
|
||
virtual Execution* onCreate(const std::vector<Tensor*>& inputs, const std::vector<Tensor*>& outputs, const MNN::Op* op) override;
|
||
```
|
||
|
||
可以在方法内根据op类型创建，但更建议提供注册接口：
|
||
```cpp
|
||
class XPUBackend final : public Backend {
|
||
// ...
|
||
class Creator {
|
||
public:
|
||
/**
|
||
* @brief create execution for given input, op on metal backend.
|
||
* @param inputs given input tensors.
|
||
* @param op given op.
|
||
* @param backend metal backend.
|
||
* @return created execution if supported, NULL otherwise.
|
||
*/
|
||
virtual Execution *onCreate(const std::vector<Tensor *> &inputs, const MNN::Op *op,
|
||
Backend *backend) const = 0;
|
||
};
|
||
/**
|
||
* @brief register creator for given op type.
|
||
* @param type given op type.
|
||
* @param creator registering creator.
|
||
*/
|
||
static void addCreator(OpType type, Creator *creator);
|
||
// ...
|
||
};
|
||
template <class T>
|
||
class XPUCreatorRegister {
|
||
public:
|
||
/**
|
||
* @brief initializer. register T creator for given op type.
|
||
* @param type given op type.
|
||
*/
|
||
XPUCreatorRegister(OpType type) {
|
||
T *test = new T;
|
||
XPUBackend::addCreator(type, test);
|
||
}
|
||
};
|
||
```
|
||
|
||
这样，Op Execution中，就可以通过注册追加Op类型：
|
||
```cpp
|
||
class XPUPoolingCreator : public XPUBackend::Creator {
|
||
public:
|
||
virtual Execution *onCreate(const std::vector<Tensor *> &inputs, const MNN::Op *op, Backend *backend) const {
|
||
return new XPUPooling(backend, op->main_as_Pool());
|
||
}
|
||
};
|
||
static XPUCreatorRegister<XPUPoolingCreator> __reg(OpType_Pooling);
|
||
```
|
||
|
||
## 内存管理
|
||
Backend通过`onAcquire`创建`MemObj`内存对象，定义其析构函数以便为tensor释放内存。内存有三种存储模式：`STATIC`内存不复用，一般用于op常量存储；`DYNAMIC`内存可复用，一般用于变量存储；`DYNAMIC_SEPERATE`内存在pipeline间可复用，一般用于pipeline常量存储。
|
||
|
||
```cpp
|
||
/** backend buffer storage type */
|
||
enum StorageType {
|
||
/**
|
||
use NOT reusable memory.
|
||
- allocates memory when `onAcquireBuffer` is called.
|
||
- releases memory when `onReleaseBuffer` is called or when the backend is deleted.
|
||
- do NOTHING when `onClearBuffer` is called.
|
||
*/
|
||
STATIC,
|
||
/**
|
||
use reusable memory.
|
||
- allocates or reuses memory when `onAcquireBuffer` is called. prefers reusing.
|
||
- collects memory for reuse when `onReleaseBuffer` is called
|
||
- releases memory when `onClearBuffer` is called or when the backend is deleted.
|
||
*/
|
||
DYNAMIC,
|
||
/**
|
||
use NOT reusable memory.
|
||
- allocates memory when `onAcquireBuffer` is called.
|
||
- do NOTHING when `onReleaseBuffer` is called.
|
||
- releases memory when `onClearBuffer` is called or when the backend is deleted.
|
||
*/
|
||
DYNAMIC_SEPERATE
|
||
};
|
||
/**
|
||
* @brief allocate buffer of tensor for given storage type.
|
||
* @param tensor buffer provider.
|
||
* @param storageType buffer storage type.
|
||
* @return MemObj for release, if failed, return nullptr.
|
||
*/
|
||
virtual MemObj* onAcquire(const Tensor* tensor, StorageType storageType) = 0;
|
||
```
|
||
|
||
Backend在调用`onClearBuffer`时，需要释放所有`DYNAMIC`和`DYNAMIC_SEPERATE`存储模式的内存：
|
||
```cpp
|
||
/**
|
||
* @brief clear all dynamic buffers.
|
||
* @return success or not.
|
||
*/
|
||
virtual bool onClearBuffer() = 0;
|
||
```
|
||
|
||
此外，backend还需要负责tensor数据的拷贝：
|
||
```cpp
|
||
/**
|
||
* @brief copy buffer from tensor to tensor.
|
||
* @param srcTensor source buffer provider.
|
||
* @param dstTensor dest buffer provider.
|
||
*/
|
||
virtual void onCopyBuffer(const Tensor* srcTensor, const Tensor* dstTensor) const = 0;
|
||
```
|
||
|
||
**拷贝可能在backend内部，也可能在backend与CPU backend之间。**
|
||
**拷贝需要处理Tensor间的布局转换，相同布局时，可以直接拷贝数据；不同布局，如**`**NHWC**`**和**`**NC4HW4**`**，则一般需要做特殊转换。**
|
||
|
||
## Pipeline回调
|
||
Backend在pipeline执行的各个周期都会收到回调，`onResizeBegin`和`onResizeEnd`在调整内存分配前后调用（op的`onResize`会在此间调用）；`onExecuteBegin`和`onExecuteEnd`在op执行前后调用（op的`onExecute`会在此间调用）；`onWaitFinish`相对特殊，由用户主动调用，异步执行的pipeline需要同步等待完成。
|
||
```cpp
|
||
/**
|
||
* @brief callback before resize ops.
|
||
*/
|
||
virtual void onResizeBegin();
|
||
/**
|
||
* @brief callback after resize ops.
|
||
*/
|
||
virtual ErrorCode onResizeEnd();
|
||
/**
|
||
* @brief callback before executing ops.
|
||
*/
|
||
virtual void onExecuteBegin() const = 0;
|
||
/**
|
||
* @brief callback after executing ops.
|
||
*/
|
||
virtual void onExecuteEnd() const = 0;
|
||
|
||
```
|
||
|
||
## Runtime（运行时）
|
||
对于使用同一种后端，且存在先后顺序，不会同时运行的模型，MNN提供机制使其共享部分计算资源，比如线程池，内存池等等。
|
||
这部分计算资源使用Runtime存储。而Backend则由Runtime创建
|
||
|
||
### CompileType
|
||
|
||
Runtime 可以通过指定 CompileType ，决定 MNN 是否跳过几何计算步骤：
|
||
|
||
```
|
||
enum CompilerType {
|
||
// 部分执行几何计算，分解形变算子，但不分解 BatchMatMul / Gather 等算子
|
||
Compiler_Geometry = 0,
|
||
|
||
// 完全跳过几何计算步骤，直接使用原始算子
|
||
Compiler_Origin = 1,
|
||
|
||
// 完全执行几何计算，仅此模式下，可以在算子不支持时自动回退到CPU计算
|
||
Compiler_Loop = 2,
|
||
};
|
||
```
|
||
|
||
### 实现Runtime
|
||
Runtime主要实现如下接口：
|
||
|
||
```
|
||
virtual Backend* onCreate(const BackendConfig* config = nullptr, Backend* origin = nullptr) const = 0;
|
||
|
||
/**
|
||
@brief reset runtime
|
||
*/
|
||
virtual void onReset(int numberThread, const BackendConfig* config, bool full) {
|
||
// Do nothing
|
||
}
|
||
|
||
/**
|
||
@brief clear unuseful resource
|
||
@param level clear level: 0 - 100, bigger mean clear more, smaller mean cache more
|
||
*/
|
||
virtual void onGabageCollect(int level) = 0;
|
||
|
||
```
|
||
|
||
- onCreate ：创建 Backend
|
||
- onReset ：重设默认配置
|
||
- onGabageCollect ：清理资源以节省内存
|
||
|
||
|
||
### 注册Runtime
|
||
注册方法中调用`MNNInsertExtraRuntimeCreator`就可以完成Runtime的注册，这里的注册方法需要在Backend.cpp中声明并调用：
|
||
```cpp
|
||
class XPURuntimeCreator : public RuntimeCreator {
|
||
virtual Runtime* onCreate(const Backend::Info &info) const {
|
||
return new XPURuntime;
|
||
}
|
||
};
|
||
void registerXPURuntimeCreator() {
|
||
MNNInsertExtraBackendCreator(MNN_FORWARD_XPU, new XPURuntimeCreator);
|
||
};
|
||
```
|
||
|
||
使用cmake编译时，完成代码修改后，也需要相应修改CMakeLists.txt。
|