Helper functions

Helper functions#

hipSPARSELt: Helper functions

Functions
hipsparseStatus_t	hipsparseLtSpMMAPrune (const hipsparseLtHandle_t handle, const hipsparseLtMatmulDescriptor_t matmulDescr, const void d_in, void d_out, hipsparseLtPruneAlg_t pruneAlg, hipStream_t stream)
	Prune a dense matrix. More...

hipsparseStatus_t	hipsparseLtSpMMAPruneCheck (const hipsparseLtHandle_t handle, const hipsparseLtMatmulDescriptor_t matmulDescr, const void d_in, int d_valid, hipStream_t stream)
	Check the correctness of the pruning structure for a given matrix. More...

hipsparseStatus_t	hipsparseLtSpMMAPrune2 (const hipsparseLtHandle_t handle, const hipsparseLtMatDescriptor_t sparseMatDescr, int isSparseA, hipsparseOperation_t op, const void d_in, void d_out, hipsparseLtPruneAlg_t pruneAlg, hipStream_t stream)
	Prune a dense matrix. More...

hipsparseStatus_t	hipsparseLtSpMMAPruneCheck2 (const hipsparseLtHandle_t handle, const hipsparseLtMatDescriptor_t sparseMatDescr, int isSparseA, hipsparseOperation_t op, const void d_in, int d_valid, hipStream_t stream)
	Check the correctness of the pruning structure for a given matrix. More...

hipsparseStatus_t	hipsparseLtSpMMACompressedSize (const hipsparseLtHandle_t handle, const hipsparseLtMatmulPlan_t plan, size_t compressedSize, size_t compressBufferSize)
	Provide the size of the compressed matrix. More...

hipsparseStatus_t	hipsparseLtSpMMACompress (const hipsparseLtHandle_t handle, const hipsparseLtMatmulPlan_t plan, const void d_dense, void d_compressed, void *d_compressBuffer, hipStream_t stream)
	Compress a dense matrix to structured matrix. More...

hipsparseStatus_t	hipsparseLtSpMMACompressedSize2 (const hipsparseLtHandle_t handle, const hipsparseLtMatDescriptor_t sparseMatDescr, size_t compressedSize, size_t compressBufferSize)
	Provide the size of the compressed matrix. More...

hipsparseStatus_t	hipsparseLtSpMMACompress2 (const hipsparseLtHandle_t handle, const hipsparseLtMatDescriptor_t sparseMatDescr, int isSparseA, hipsparseOperation_t op, const void d_dense, void d_compressed, void *d_compressBuffer, hipStream_t stream)
	Compress a dense matrix to structured matrix. More...

Detailed Description

Required for subsequent library calls.

Function Documentation

◆ hipsparseLtSpMMACompress()

hipsparseStatus_t hipsparseLtSpMMACompress	(	const hipsparseLtHandle_t *	handle,
		const hipsparseLtMatmulPlan_t *	plan,
		const void *	d_dense,
		void *	d_compressed,
		void *	d_compressBuffer,
		hipStream_t	stream
	)

Compress a dense matrix to structured matrix.

hipsparseLtSpMMACompress compresses the dense matrix d_dense. The compressed matrix is intended to be used as the first/second operand A/B in the hipsparseLtMatmul() function.

Parameters

[in]	handle	handle to the hipsparselt library context queue.
[in]	plan	matrix multiplication plan descriptor.
[in]	d_dense	pointer to the dense matrix.
[out]	d_compressed	compressed matrix and metadata.
[out]	d_compressBuffer	temporary buffer for the compression.
[in]	stream	HIP stream for the computation.

Return values

HIPSPARSE_STATUS_SUCCESS	the operation completed successfully.
HIPSPARSE_STATUS_INVALID_VALUE	`handle`, `plan`, `d_dense`, or `d_compressed` is invalid.
HIPSPARSE_STATUS_NOT_SUPPORTED	the problem is not supported.

◆ hipsparseLtSpMMACompress2()

hipsparseStatus_t hipsparseLtSpMMACompress2	(	const hipsparseLtHandle_t *	handle,
		const hipsparseLtMatDescriptor_t *	sparseMatDescr,
		int	isSparseA,
		hipsparseOperation_t	op,
		const void *	d_dense,
		void *	d_compressed,
		void *	d_compressBuffer,
		hipStream_t	stream
	)

Compress a dense matrix to structured matrix.

hipsparseLtSpMMACompress2 compresses the dense matrix d_dense. The compressed matrix is intended to be used as the first/second operand A/B in the hipsparseLtMatmul() function.

Parameters

[in]	handle	handle to the hipsparselt library context queue.
[in]	sparseMatDescr	structured (sparse) matrix descriptor.
[in]	isSparseA	specify whether the structured (or sparse) matrix is in the first position (matA or matB).
[in]	op	operation that will be applied to the structured (or sparse) matrix in the multiplication.
[in]	d_dense	pointer to the dense matrix.
[out]	d_compressed	compressed matrix and metadata.
[out]	d_compressBuffer	temporary buffer for the compression.
[in]	stream	HIP stream for the computation.

Return values

HIPSPARSE_STATUS_SUCCESS	the operation completed successfully.
HIPSPARSE_STATUS_INVALID_VALUE	`handle`, `sparseMatDescr`, `op`, `d_dense`, or `d_compressed` is invalid.
HIPSPARSE_STATUS_NOT_SUPPORTED	the problem is not supported.

◆ hipsparseLtSpMMACompressedSize()

hipsparseStatus_t hipsparseLtSpMMACompressedSize	(	const hipsparseLtHandle_t *	handle,
		const hipsparseLtMatmulPlan_t *	plan,
		size_t *	compressedSize,
		size_t *	compressBufferSize
	)

Provide the size of the compressed matrix.

hipsparseLtSpMMACompressedSize provides the size of the compressed matrix to be allocated before calling hipsparseLtSpMMACompress() or hipsparseLtSpMMACompress2().

Parameters

[in]	handle	hipsparselt library handle.
[in]	plan	matrix multiplication plan descriptor.
[out]	compressedSize	size in bytes of the compressed matrix.
[out]	compressBufferSize	size in bytes for the buffer needed for the matrix compression.

Return values

HIPSPARSE_STATUS_SUCCESS	the operation completed successfully.
HIPSPARSE_STATUS_INVALID_VALUE	`handle`, `plan`, `compressedSize`, or `compressBufferSize` is invalid.

◆ hipsparseLtSpMMACompressedSize2()

hipsparseStatus_t hipsparseLtSpMMACompressedSize2	(	const hipsparseLtHandle_t *	handle,
		const hipsparseLtMatDescriptor_t *	sparseMatDescr,
		size_t *	compressedSize,
		size_t *	compressBufferSize
	)

Provide the size of the compressed matrix.

hipsparseLtSpMMACompressedSize2 provides the size of the compressed matrix to be allocated before calling hipsparseLtSpMMACompress or hipsparseLtSpMMACompress2.

Parameters

[in]	handle	hipsparselt library handle.
[in]	sparseMatDescr	structured (sparse) matrix descriptor.
[out]	compressedSize	size in bytes of the compressed matrix.
[out]	compressBufferSize	size in bytes for the buffer needed for the matrix compression.

Return values

HIPSPARSE_STATUS_SUCCESS	the operation completed successfully.
HIPSPARSE_STATUS_NOT_INITIALIZED	`handle`, `sparseMatDescr`, `compressedSize`, or `compressBufferSize` is invalid.
HIPSPARSE_STATUS_NOT_SUPPORTED	the problem is not supported.

◆ hipsparseLtSpMMAPrune()

hipsparseStatus_t hipsparseLtSpMMAPrune	(	const hipsparseLtHandle_t *	handle,
		const hipsparseLtMatmulDescriptor_t *	matmulDescr,
		const void *	d_in,
		void *	d_out,
		hipsparseLtPruneAlg_t	pruneAlg,
		hipStream_t	stream
	)

Prune a dense matrix.

hipsparseLtSpMMAPrune prunes the dense matrix d_in according to the specified algorithm pruneAlg, which can be HIPSPARSELT_PRUNE_SPMMA_TILE or HIPSPARSELT_PRUNE_SPMMA_STRIP.

Note: The function requires no extra storage. It supports asynchronous execution with respect to stream.

Parameters

[in]	handle	hipsparselt library handle.
[in]	matmulDescr	matrix multiplication descriptor.
[in]	d_in	pointer to the dense matrix.
[out]	d_out	pointer to the pruned matrix.
[in]	pruneAlg	pruning algorithm.
[in]	stream	HIP stream for the computation.

Return values

HIPSPARSE_STATUS_SUCCESS	the operation completed successfully.
HIPSPARSE_STATUS_INVALID_VALUE	`handle`, `matmulDescr`, `d_in`, or `d_out` is invalid.

◆ hipsparseLtSpMMAPrune2()

hipsparseStatus_t hipsparseLtSpMMAPrune2	(	const hipsparseLtHandle_t *	handle,
		const hipsparseLtMatDescriptor_t *	sparseMatDescr,
		int	isSparseA,
		hipsparseOperation_t	op,
		const void *	d_in,
		void *	d_out,
		hipsparseLtPruneAlg_t	pruneAlg,
		hipStream_t	stream
	)

Prune a dense matrix.

hipsparseLtSpMMAPrune2 prunes the dense matrix d_in according to the specified algorithm pruneAlg, which can be HIPSPARSELT_PRUNE_SPMMA_TILE or HIPSPARSELT_PRUNE_SPMMA_STRIP.

Note: The function requires no extra storage. It supports asynchronous execution with respect to stream.

Parameters

[in]	handle	hipsparselt library handle.
[in]	sparseMatDescr	structured (sparse) matrix descriptor.
[in]	isSparseA	specify if the structured (or sparse) matrix is in the first position (matA or matB). (It currently only supports matA.)
[in]	op	operation that will be applied to the structured (or sparse) matrix in the multiplication.
[in]	d_in	pointer to the dense matrix.
[out]	d_out	pointer to the pruned matrix.
[in]	pruneAlg	pruning algorithm.
[in]	stream	HIP stream for the computation.

Return values

HIPSPARSE_STATUS_SUCCESS	the operation completed successfully.
HIPSPARSE_STATUS_INVALID_VALUE	`handle`, `sparseMatDescr`, `op`, `d_in`, or `d_out` is invalid.
HIPSPARSE_STATUS_NOT_SUPPORTED	the problem is not supported.

◆ hipsparseLtSpMMAPruneCheck()

hipsparseStatus_t hipsparseLtSpMMAPruneCheck	(	const hipsparseLtHandle_t *	handle,
		const hipsparseLtMatmulDescriptor_t *	matmulDescr,
		const void *	d_in,
		int *	d_valid,
		hipStream_t	stream
	)

Check the correctness of the pruning structure for a given matrix.

hipsparseLtSpMMAPruneCheck checks the correctness of the pruning structure for a given matrix. Contents in the provided matrix must have a sparsity of 2:4.

Parameters

[in]	handle	hipsparselt library handle.
[in]	matmulDescr	matrix multiplication descriptor.
[in]	d_in	pointer to the matrix to check.
[out]	d_valid	validation results (0 is correct, and 1 is incorrect).
[in]	stream	HIP stream for the computation.

Return values

HIPSPARSE_STATUS_SUCCESS	the operation completed successfully.
HIPSPARSE_STATUS_INVALID_VALUE	`handle`, `matmulDescr`, `d_in`, or `d_valid` is invalid.

◆ hipsparseLtSpMMAPruneCheck2()

hipsparseStatus_t hipsparseLtSpMMAPruneCheck2	(	const hipsparseLtHandle_t *	handle,
		const hipsparseLtMatDescriptor_t *	sparseMatDescr,
		int	isSparseA,
		hipsparseOperation_t	op,
		const void *	d_in,
		int *	d_valid,
		hipStream_t	stream
	)

Check the correctness of the pruning structure for a given matrix.

hipsparseLtSpMMAPruneCheck2 checks the correctness of the pruning structure for a given matrix. Contents in the provided matrix must have a sparsity of 2:4.

Parameters

[in]	handle	hipsparselt library handle.
[in]	sparseMatDescr	structured (sparse) matrix descriptor.
[in]	isSparseA	specify if the structured (or sparse) matrix is in the first position (matA or matB). (The HIP backend only supports matA.)
[in]	op	operation that will be applied to the structured (or sparse) matrix in the multiplication.
[in]	d_in	pointer to the matrix to check.
[out]	d_valid	validation results (0 is correct, and 1 is incorrect).
[in]	stream	HIP stream for the computation.

Return values

HIPSPARSE_STATUS_SUCCESS	the operation completed successfully.
HIPSPARSE_STATUS_INVALID_VALUE	`handle`, `sparseMatDescr`, `op`, `d_in`, or `d_valid` is invalid.
HIPSPARSE_STATUS_NOT_SUPPORTED	the problem is not supported.

Helper functions

Helper functions#

Functions

Detailed Description

Function Documentation

◆ hipsparseLtSpMMACompress()

◆ hipsparseLtSpMMACompress2()

◆ hipsparseLtSpMMACompressedSize()

◆ hipsparseLtSpMMACompressedSize2()

◆ hipsparseLtSpMMAPrune()

◆ hipsparseLtSpMMAPrune2()

◆ hipsparseLtSpMMAPruneCheck()

◆ hipsparseLtSpMMAPruneCheck2()