Dataset

oneMKL - Data Parallel C++ Developer Reference

Download PDF

ID 772045

Date 7/13/2023

Version

Public

A newer version of this document is available. Customers should click here to go to the newest version.

Dataset

The oneapi::mkl::stats::dataset structure consolidates information of a multi-dimensional dataset.

API

Syntax

namespace oneapi::mkl::stats {
  template<layout ObservationsLayout = layout::row_major,
  typename Type = float*>

namespace oneapi::mkl::stats {
  struct dataset {}
}

Buffer API specialization

namespace oneapi::mkl::stats {

  template<layout ObservationsLayout, typename Type>struct dataset<ObservationsLayout, sycl::buffer<Type, 1>> {

     explicit dataset(std::int64_t n_dims_, std::int64_t n_observations_,
           sycl::buffer<Type, 1> observations_, sycl::buffer<Type, 1> weights_ = {0},
           sycl::buffer<std::int64_t, 1> indices_ = {0}) : n_dims(n_dims_), n_observations(n_observations_), observations(observations_),
           weights(weights_), indices(indices_);

 std::int64_t n_dims;
 std::int64_t n_observations;
 sycl::buffer<Type, 1> observations;
 sycl::buffer<Type, 1> weights = {0};
 sycl::buffer<std::int64_t, 1> indices = {0};
 static constexpr layout layout = ObservationsLayout;
 };
}

USM API specialization

namespace oneapi::mkl::stats {

 template< layout ObservationsLayout, typename Type>
 struct dataset<ObservationsLayout, Type*> {
      explicit dataset(std::int64_t n_dims_, std::int64_t n_observations_, Type* observations_,
           Type* weights_ = nullptr, std::int64_t* indices_ = nullptr) :
           n_dims(n_dims_), n_observations(n_observations_),
           observations(observations_),
           weights(weights_), indices(indices_);


 std::int64_t n_dims;
 std::int64_t n_observations;
 Type* observations;
 Type* weights = nullptr;
 std::int64_t* indices = nullptr;
 static constexpr layout layout = ObservationsLayout;
 };
}

Include Files

oneapi/mkl/stats.hpp

Template Parameters

`typename DataType`	Type of dataset. May be `sycl::buffer<float,1>`, `sycl::buffer<double,1>` (for buffer-based dataset) or `float`, `double` (for USM-based dataset).
`oneapi::mkl::stats::layout ObservationsLayout`	Layout of the observations matrix of the dataset. The specific values are as follows: `oneapi::mkl::stats::layout::row_majoroneapi::mkl::stats::layout::col_major`

Struct Members

Name	Type	Description
n_dims	`std::int64_t`	The number of dimensions (variables)
n_observations	`std::int64_t`	The number of observations
layout	`oneapi::mkl::stats::layout`	Characterize the matrix of observations layout (row or column major)
observations	`sycl::buffer<Type, 1> / Type*`	Matrix of observations.
weights	`sycl::buffer<Type, 1> / Type*`	Array of weights of size n_observations. Elements of the arrays are non-negative numbers. If the parameter is not specified, each observation is assigned a weight equal to 1.
indices	`sycl::buffer<std::int64_t, 1> / std::int64_t*`	Array of vector components that will be processed. Size of array is `n_dims`. If the parameter is not specified, all components of the vector are processed.

Select Your Language

Using Intel.com Search

Quick Links

Recent Searches

Advanced Search

Only search in

oneMKL - Data Parallel C++ Developer Reference

Dataset

API