cuVectorOperations< DataType > Class Template Reference

A static class for vector operations, similar to level-1 operations of the BLAS library. This class acts as a templated namespace, where all member methods are public and static. More...

#include <cu_vector_operations.h>

Static Public Member Functions
static void	copy_vector (cublasHandle_t cublas_handle, const DataType *RESTRICT input_vector, const LongIndexType vector_size, DataType *RESTRICT output_vector)
	Copies a vector to a new vector. Result is written in-place.
static void	copy_scaled_vector (cublasHandle_t cublas_handle, const DataType *RESTRICT input_vector, const LongIndexType vector_size, const DataType scale, DataType *RESTRICT output_vector)
	Scales a vector and stores to a new vector.
static void	subtract_scaled_vector (cublasHandle_t cublas_handle, const DataType *RESTRICT input_vector, const LongIndexType vector_size, const DataType scale, DataType *RESTRICT output_vector)
	Subtracts the scaled input vector from the output vector.
static DataType	inner_product (cublasHandle_t cublas_handle, const DataType *RESTRICT vector1, const DataType *RESTRICT vector2, const LongIndexType vector_size)
	Computes Euclidean inner product of two vectors.
static DataType	euclidean_norm (cublasHandle_t cublas_handle, const DataType *RESTRICT vector, const LongIndexType vector_size)
	Computes the Euclidean 2-norm of a 1D array.
static DataType	normalize_vector_in_place (cublasHandle_t cublas_handle, DataType *RESTRICT vector, const LongIndexType vector_size)
	Normalizes a vector based on Euclidean 2-norm. The result is written in-place.
static DataType	normalize_vector_and_copy (cublasHandle_t cublas_handle, const DataType *RESTRICT vector, const LongIndexType vector_size, DataType *RESTRICT output_vector)
	Normalizes a vector based on Euclidean 2-norm. The result is written into another vector.

Detailed Description

template<typename DataType>
class cuVectorOperations< DataType >

A static class for vector operations, similar to level-1 operations of the BLAS library. This class acts as a templated namespace, where all member methods are public and static.

See also

MatrixOperations

Definition at line 64 of file cu_vector_operations.h.

Member Function Documentation

copy_scaled_vector()

template<typename DataType >

void cuVectorOperations< DataType >::copy_scaled_vector ( cublasHandle_t  cublas_handle, const DataType *RESTRICT  input_vector, const LongIndexType  vector_size, const DataType  scale, DataType *RESTRICT  output_vector  )

static

Scales a vector and stores to a new vector.

Parameters

[in]	cublas_handle	The cuBLAS object handle.
[in]	input_vector	A 1D array
[in]	vector_size	Length of vector array
[in]	scale	Scale coefficient to the input vector. If this is equal to one, the function effectively becomes the same as copy_vector.
[out]	output_vector	Output vector (written in place).

Definition at line 76 of file cu_vector_operations.cu.

{
    cublasStatus_t status;
    int incx = 1;
    int incy = 1;
 
    // Copy input to output vector
    status = cublas_api::cublasXcopy(cublas_handle, vector_size, input_vector,
                                     incx, output_vector, incy);
 
    assert(status == CUBLAS_STATUS_SUCCESS);
 
    // Scale output vector
    status = cublas_api::cublasXscal(cublas_handle, vector_size, &scale,
                                     output_vector, incy);
 
    assert(status == CUBLAS_STATUS_SUCCESS);
}

References cublas_api::cublasXcopy(), and cublas_api::cublasXscal().

Referenced by cu_lanczos_tridiagonalization(), and cuVectorOperations< DataType >::normalize_vector_and_copy().

Here is the call graph for this function:

Here is the caller graph for this function:

copy_vector()

template<typename DataType >

void cuVectorOperations< DataType >::copy_vector ( cublasHandle_t  cublas_handle, const DataType *RESTRICT  input_vector, const LongIndexType  vector_size, DataType *RESTRICT  output_vector  )

static

Copies a vector to a new vector. Result is written in-place.

Parameters

[in]	cublas_handle	The cuBLAS object handle.
[in]	input_vector	A 1D array
[in]	vector_size	Length of vector array
[out]	output_vector	Output vector (written in place).

Definition at line 41 of file cu_vector_operations.cu.

{
    int incx = 1;
    int incy = 1;
 
    cublasStatus_t status = cublas_api::cublasXcopy(
            cublas_handle, vector_size, input_vector, incx, output_vector,
            incy);
 
    assert(status == CUBLAS_STATUS_SUCCESS);
}

References cublas_api::cublasXcopy().

Here is the call graph for this function:

euclidean_norm()

template<typename DataType >

DataType cuVectorOperations< DataType >::euclidean_norm ( cublasHandle_t  cublas_handle, const DataType *RESTRICT  vector, const LongIndexType  vector_size  )

static

Computes the Euclidean 2-norm of a 1D array.

Parameters

[in]	cublas_handle	The cuBLAS object handle.
[in]	vector	A pointer to 1D array
[in]	vector_size	Length of the array

Returns

Euclidean norm

Definition at line 204 of file cu_vector_operations.cu.

{
    DataType norm;
    int incx = 1;
 
    cublasStatus_t status = cublas_api::cublasXnrm2(
            cublas_handle, vector_size, vector, incx, &norm);
 
    assert(status == CUBLAS_STATUS_SUCCESS);
 
    return norm;
}

References cublas_api::cublasXnrm2().

Referenced by cu_lanczos_tridiagonalization(), cuOrthogonalization< DataType >::gram_schmidt_process(), cuVectorOperations< DataType >::normalize_vector_and_copy(), cuVectorOperations< DataType >::normalize_vector_in_place(), and cuOrthogonalization< DataType >::orthogonalize_vectors().

Here is the call graph for this function:

Here is the caller graph for this function:

inner_product()

template<typename DataType >

DataType cuVectorOperations< DataType >::inner_product ( cublasHandle_t  cublas_handle, const DataType *RESTRICT  vector1, const DataType *RESTRICT  vector2, const LongIndexType  vector_size  )

static

Computes Euclidean inner product of two vectors.

Parameters

[in]	cublas_handle	The cuBLAS object handle.
[in]	vector1	1D array
[in]	vector2	1D array
[in]	vector_size	Length of array

Returns

Inner product of two vectors.

Definition at line 169 of file cu_vector_operations.cu.

{
    DataType inner_prod;
    int incx = 1;
    int incy = 1;
 
    cublasStatus_t status = cublas_api::cublasXdot(
            cublas_handle, vector_size, vector1, incx, vector2, incy,
            &inner_prod);
 
    assert(status == CUBLAS_STATUS_SUCCESS);
 
    return inner_prod;
}

References cublas_api::cublasXdot().

Referenced by cu_lanczos_tridiagonalization(), cuOrthogonalization< DataType >::gram_schmidt_process(), and cuOrthogonalization< DataType >::orthogonalize_vectors().

Here is the call graph for this function:

Here is the caller graph for this function:

normalize_vector_and_copy()

template<typename DataType >

DataType cuVectorOperations< DataType >::normalize_vector_and_copy ( cublasHandle_t  cublas_handle, const DataType *RESTRICT  vector, const LongIndexType  vector_size, DataType *RESTRICT  output_vector  )

static

Normalizes a vector based on Euclidean 2-norm. The result is written into another vector.

Parameters

[in]	cublas_handle	The cuBLAS object handle.
[in]	vector	Input vector.
[in]	vector_size	Length of the input vector
[out]	output_vector	Output vector, which is the normalization of the input vector.

Returns

2-norm of the input vector

Definition at line 278 of file cu_vector_operations.cu.

{
    // Norm of vector
    DataType norm = cuVectorOperations<DataType>::euclidean_norm(
            cublas_handle, vector, vector_size);
 
    // Normalize to output
    DataType scale = cu_arithmetics::div(
            cu_arithmetics::cast<double, DataType>(1.0),
            norm);
    cuVectorOperations<DataType>::copy_scaled_vector(cublas_handle, vector,
                                                     vector_size, scale,
                                                     output_vector);
 
    return norm;
}

References cu_arithmetics::abs(), cuVectorOperations< DataType >::copy_scaled_vector(), cu_arithmetics::div(), and cuVectorOperations< DataType >::euclidean_norm().

Here is the call graph for this function:

normalize_vector_in_place()

template<typename DataType >

DataType cuVectorOperations< DataType >::normalize_vector_in_place ( cublasHandle_t  cublas_handle, DataType *RESTRICT  vector, const LongIndexType  vector_size  )

static

Normalizes a vector based on Euclidean 2-norm. The result is written in-place.

Parameters

[in]	cublas_handle	The cuBLAS object handle.
[in,out]	vector	Input vector to be normalized in-place.
[in]	vector_size	Length of the input vector

Returns

2-Norm of the input vector (before normalization)

Definition at line 237 of file cu_vector_operations.cu.

{
    // Norm of vector
    DataType norm = cuVectorOperations<DataType>::euclidean_norm(
            cublas_handle, vector, vector_size);
 
    // Normalize in place
    DataType scale = cu_arithmetics::div(
            cu_arithmetics::cast<double, DataType>(1.0),
            norm);
    int incx = 1;
    cublasStatus_t status = cublas_api::cublasXscal(
            cublas_handle, vector_size, &scale, vector, incx);
 
    assert(status == CUBLAS_STATUS_SUCCESS);
 
    return norm;
}

References cu_arithmetics::abs(), cublas_api::cublasXscal(), cu_arithmetics::div(), and cuVectorOperations< DataType >::euclidean_norm().

Referenced by cu_golub_kahn_bidiagonalization().

Here is the call graph for this function:

Here is the caller graph for this function:

subtract_scaled_vector()

template<typename DataType >

void cuVectorOperations< DataType >::subtract_scaled_vector ( cublasHandle_t  cublas_handle, const DataType *RESTRICT  input_vector, const LongIndexType  vector_size, const DataType  scale, DataType *RESTRICT  output_vector  )

static

Subtracts the scaled input vector from the output vector.

Performs the following operation:

\[ \boldsymbol{b} = \boldsymbol{b} - c \boldsymbol{a}, \]

where

• \( \boldsymbol{a} \) is the input vector,
• \( c \) is a scalar scale to the input vector, and
• \( \boldsymbol{b} \) is the output vector that is written in-place.

Parameters

[in]	cublas_handle	The cuBLAS object handle.
[in]	input_vector	A 1D array
[in]	vector_size	Length of vector array
[in]	scale	Scale coefficient to the input vector.
[in,out]	output_vector	Output vector (written in place).

Definition at line 128 of file cu_vector_operations.cu.

{
    DataType zero = 0.0;
    if (cu_arithmetics::is_equal(scale, zero))
    {
        return;
    }
 
    int incx = 1;
    int incy = 1;
 
    DataType neg_scale = -scale;
    cublasStatus_t status = cublas_api::cublasXaxpy(
            cublas_handle, vector_size, &neg_scale, input_vector, incx,
            output_vector, incy);
 
    assert(status == CUBLAS_STATUS_SUCCESS);
}

References cublas_api::cublasXaxpy(), and cu_arithmetics::is_equal().

Referenced by cuAffineMatrixFunction< DataType >::_add_scaled_vector(), cu_golub_kahn_bidiagonalization(), cu_lanczos_tridiagonalization(), cuOrthogonalization< DataType >::gram_schmidt_process(), and cuOrthogonalization< DataType >::orthogonalize_vectors().

Here is the call graph for this function:

Here is the caller graph for this function:

---

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

• /home/runner/work/imate/imate/imate/_cu_basic_algebra/cu_vector_operations.h
• /home/runner/work/imate/imate/imate/_cu_basic_algebra/cu_vector_operations.cu