Managed reference to PostprocNode. More...

#include <postproc.h>

Inheritance diagram for tvm::meta_schedule::Postproc:

Collaboration diagram for tvm::meta_schedule::Postproc:

Public Types
using	FInitializeWithTuneContext = ffi::TypedFunction< void(const TuneContext &)>
	The function type of `InitializeWithTuneContext` method. More...

using	FApply = ffi::TypedFunction< bool(const tir::Schedule &)>
	Apply a postprocessor to the given schedule. More...

using	FClone = ffi::TypedFunction< Postproc()>
	Clone the postprocessor. More...

using	FAsString = ffi::TypedFunction< String()>
	Get the postprocessor function as string with name. More...

Public Member Functions
	TVM_DEFINE_MUTABLE_OBJECT_REF_METHODS (Postproc, ObjectRef, PostprocNode)

Static Public Member Functions
static Postproc	PyPostproc (FInitializeWithTuneContext f_initialize_with_tune_context, FApply f_apply, FClone f_clone, FAsString f_as_string)
	Create a postprocessor with customized methods on the python-side. More...

static Postproc	DisallowDynamicLoop ()
	Create a postprocessor that checks if all loops are static. More...

static Postproc	DisallowAsyncStridedMemCopy ()
	Create a postprocessor that checks if all async mem copies are not strided. More...

static Postproc	RewriteCooperativeFetch ()
	Create a postprocessor that rewrites the cooperative fetch annotation to actual vectorized cooperative fetching in loop bindings. More...

static Postproc	RewriteParallelVectorizeUnroll ()
	Creates a postprocessor that applies parallelization, vectorization and auto unrolling according to the annotation of each block. More...

static Postproc	RewriteReductionBlock ()
	Create a postprocessor that rewrites reduction block by moving the init block out. More...

static Postproc	RewriteUnboundBlock (int max_threadblocks)
	Create a postprocessor that adds thread binding to unbound blocks. More...

static Postproc	RewriteTensorize (bool vectorize_init_loop=false)
	Create a postprocessor that applies tensorization to annotated blocks. More...

static Postproc	VerifyGPUCode ()
	Creates a postprocessor that verifies if the GPU code is correct. More...

static Postproc	VerifyVTCMLimit ()
	Verifies that the VTCM usage of a given schedule is within the provided limit. More...

static Postproc	RewriteLayout ()
	Creates a postprocessor that rewrites the layout of input tensor. More...

static Array< Postproc, void >	DefaultLLVM ()
	Create default postprocessors for LLVM. More...

static Array< Postproc, void >	DefaultCPUTensorization ()
	Create default postprocessors for x86 (AVX512 and VNNI) More...

static Array< Postproc, void >	DefaultCUDA ()
	Create default postprocessors for CUDA. More...

static Array< Postproc, void >	DefaultCUDATensorCore ()
	Create default postprocessors for CUDA with TensorCore. More...

static Array< Postproc, void >	DefaultHexagon ()
	Create default postprocessors for Hexagon. More...

Detailed Description

Managed reference to PostprocNode.

See also: PostprocNode

Member Typedef Documentation

◆ FApply

using tvm::meta_schedule::Postproc::FApply = ffi::TypedFunction<bool(const tir::Schedule&)>

Apply a postprocessor to the given schedule.

Parameters

sch	The schedule to be post processed.

Returns: Whether the postprocessor was successfully applied.

◆ FAsString

using tvm::meta_schedule::Postproc::FAsString = ffi::TypedFunction<String()>

Get the postprocessor function as string with name.

Returns: The string of the postprocessor function.

◆ FClone

using tvm::meta_schedule::Postproc::FClone = ffi::TypedFunction<Postproc()>

Clone the postprocessor.

Returns: The cloned postprocessor.

◆ FInitializeWithTuneContext

using tvm::meta_schedule::Postproc::FInitializeWithTuneContext = ffi::TypedFunction<void(const TuneContext&)>

The function type of InitializeWithTuneContext method.

Parameters

context The tuning context for initialization.

Member Function Documentation

◆ DefaultCPUTensorization()

static Array<Postproc, void> tvm::meta_schedule::Postproc::DefaultCPUTensorization ( )

static

Create default postprocessors for x86 (AVX512 and VNNI)

◆ DefaultCUDA()

static Array<Postproc, void> tvm::meta_schedule::Postproc::DefaultCUDA ( )

static

Create default postprocessors for CUDA.

◆ DefaultCUDATensorCore()

static Array<Postproc, void> tvm::meta_schedule::Postproc::DefaultCUDATensorCore ( )

static

Create default postprocessors for CUDA with TensorCore.

◆ DefaultHexagon()

static Array<Postproc, void> tvm::meta_schedule::Postproc::DefaultHexagon ( )

static

Create default postprocessors for Hexagon.

◆ DefaultLLVM()

static Array<Postproc, void> tvm::meta_schedule::Postproc::DefaultLLVM ( )

static

Create default postprocessors for LLVM.

◆ DisallowAsyncStridedMemCopy()

static Postproc tvm::meta_schedule::Postproc::DisallowAsyncStridedMemCopy ( )

static

Create a postprocessor that checks if all async mem copies are not strided.

Returns: The postprocessor created

◆ DisallowDynamicLoop()

static Postproc tvm::meta_schedule::Postproc::DisallowDynamicLoop ( )

static

Create a postprocessor that checks if all loops are static.

Returns: The postprocessor created

◆ PyPostproc()

static Postproc tvm::meta_schedule::Postproc::PyPostproc	(	FInitializeWithTuneContext	f_initialize_with_tune_context,
		FApply	f_apply,
		FClone	f_clone,
		FAsString	f_as_string
	)

static

Create a postprocessor with customized methods on the python-side.

Parameters

f_initialize_with_tune_context	The packed function of `InitializeWithTuneContext`.
f_apply	The packed function of `Apply`.
f_clone	The packed function of `Clone`.
f_as_string	The packed function of `AsString`.

Returns: The postprocessor created.

◆ RewriteCooperativeFetch()

static Postproc tvm::meta_schedule::Postproc::RewriteCooperativeFetch ( )

static

Create a postprocessor that rewrites the cooperative fetch annotation to actual vectorized cooperative fetching in loop bindings.

Returns: The postprocessor created.

◆ RewriteLayout()

static Postproc tvm::meta_schedule::Postproc::RewriteLayout ( )

static

Creates a postprocessor that rewrites the layout of input tensor.

Note: Weight layout rewrite is supported so far, activation layout rewrite will be added.

Returns: The postprocessor created

◆ RewriteParallelVectorizeUnroll()

static Postproc tvm::meta_schedule::Postproc::RewriteParallelVectorizeUnroll ( )

static

Creates a postprocessor that applies parallelization, vectorization and auto unrolling according to the annotation of each block.

Returns: The postprocessor created

◆ RewriteReductionBlock()

static Postproc tvm::meta_schedule::Postproc::RewriteReductionBlock ( )

static

Create a postprocessor that rewrites reduction block by moving the init block out.

Returns: The postprocessor created.

◆ RewriteTensorize()

static Postproc tvm::meta_schedule::Postproc::RewriteTensorize ( bool vectorize_init_loop = false )

static

Create a postprocessor that applies tensorization to annotated blocks.

Parameters

vectorize_init_loop Whether or not vectorize the initialization loop produced by DecomposeReduction

Returns: The postprocessor created.

◆ RewriteUnboundBlock()

static Postproc tvm::meta_schedule::Postproc::RewriteUnboundBlock ( int max_threadblocks )

static

Create a postprocessor that adds thread binding to unbound blocks.

Parameters

max_threadblocks The max number of threadblocks in the cuda device.

Returns: The postprocessor created.

◆ TVM_DEFINE_MUTABLE_OBJECT_REF_METHODS()

tvm::meta_schedule::Postproc::TVM_DEFINE_MUTABLE_OBJECT_REF_METHODS	(	Postproc	,
		ObjectRef	,
		PostprocNode
	)

◆ VerifyGPUCode()

static Postproc tvm::meta_schedule::Postproc::VerifyGPUCode ( )

static

Creates a postprocessor that verifies if the GPU code is correct.

Returns: The postprocessor created

◆ VerifyVTCMLimit()

static Postproc tvm::meta_schedule::Postproc::VerifyVTCMLimit ( )

static

Verifies that the VTCM usage of a given schedule is within the provided limit.

Returns: The postprocessor created

The documentation for this class was generated from the following file:

include/tvm/meta_schedule/postproc.h

Public Types

Public Member Functions

Static Public Member Functions

Detailed Description

Member Typedef Documentation

◆ FApply

◆ FAsString

◆ FClone

◆ FInitializeWithTuneContext

Member Function Documentation

◆ DefaultCPUTensorization()

◆ DefaultCUDA()

◆ DefaultCUDATensorCore()

◆ DefaultHexagon()

◆ DefaultLLVM()

◆ DisallowAsyncStridedMemCopy()

◆ DisallowDynamicLoop()

◆ PyPostproc()

◆ RewriteCooperativeFetch()

◆ RewriteLayout()

◆ RewriteParallelVectorizeUnroll()

◆ RewriteReductionBlock()

◆ RewriteTensorize()

◆ RewriteUnboundBlock()

◆ TVM_DEFINE_MUTABLE_OBJECT_REF_METHODS()

◆ VerifyGPUCode()

◆ VerifyVTCMLimit()