0855805ce2
This commits adds an implementation using the C++ parallel STL. The Makefile uses the NVIDIA HPC SDK `nvc++` compiler with the `-stdpar` flag. Tested using the NVIDIA HPC SDK 20.9.
39 lines
874 B
C++
39 lines
874 B
C++
// Copyright (c) 2020, NVIDIA CORPORATION. All rights reserved.
|
|
//
|
|
// For full license terms please see the LICENSE file distributed with this
|
|
// source code
|
|
|
|
#pragma once
|
|
|
|
#include <iostream>
|
|
#include <stdexcept>
|
|
#include "Stream.h"
|
|
|
|
#define IMPLEMENTATION_STRING "STD"
|
|
|
|
template <class T>
|
|
class STDStream : public Stream<T>
|
|
{
|
|
protected:
|
|
// Size of arrays
|
|
unsigned int array_size;
|
|
|
|
// Device side pointers
|
|
T *a;
|
|
T *b;
|
|
T *c;
|
|
|
|
public:
|
|
STDStream(const unsigned int, T*, T*, T*, int);
|
|
~STDStream() = default;
|
|
|
|
virtual void copy() override;
|
|
virtual void add() override;
|
|
virtual void mul() override;
|
|
virtual void triad() override;
|
|
virtual T dot() override;
|
|
|
|
virtual void init_arrays(T initA, T initB, T initC) override;
|
|
virtual void read_arrays(std::vector<T>& a, std::vector<T>& b, std::vector<T>& c) override;
|
|
};
|