test/conformance/conformance_parallel_scan.cpp

/*
    Copyright (c) 2020-2023 Intel Corporation

    Licensed under the Apache License, Version 2.0 (the "License");
    you may not use this file except in compliance with the License.
    You may obtain a copy of the License at

        http://www.apache.org/licenses/LICENSE-2.0

    Unless required by applicable law or agreed to in writing, software
    distributed under the License is distributed on an "AS IS" BASIS,
    WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
    See the License for the specific language governing permissions and
    limitations under the License.
*/

#if __INTEL_COMPILER && _MSC_VER
#pragma warning(disable : 2586) // decorated name length exceeded, name was truncated
#endif

#define DOCTEST_CONFIG_SUPER_FAST_ASSERTS
#include "common/test.h"
#include "common/test_invoke.h"

#include "oneapi/tbb/parallel_scan.h"

#include <vector>

//! \file conformance_parallel_scan.cpp
//! \brief Test for [algorithms.parallel_scan] specification

constexpr std::size_t size = 1000;

template<typename T, typename Op>
class Body {
    const T identity;
    T sum;
    std::vector<T>& y;
    const std::vector<T>& z;
public:
    Body( const std::vector<T>& z_, std::vector<T>& y_, T id ) : identity(id), sum(id), y(y_), z(z_) {}
    T get_sum() const { return sum; }

    template<typename Tag>
    void operator()( const oneapi::tbb::blocked_range<std::size_t>& r, Tag ) {
        T temp = sum;
        for(std::size_t i=r.begin(); i<r.end(); ++i ) {
            temp = Op()(temp, z[i]);
            if( Tag::is_final_scan() )
                y[i] = temp;
        }
        sum = temp;
    }
    Body( Body& b, oneapi::tbb::split ): identity(b.identity), sum(b.identity), y(b.y), z(b.z) {}
    void reverse_join( Body& a ) { sum = Op()(a.sum, sum); }
    void assign( Body& b ) { sum = b.sum; }
};

class default_partitioner_tag{};

template<typename Partitioner>
struct parallel_scan_wrapper{
    template<typename... Args>
    void operator()(Args&&... args) {
    oneapi::tbb::parallel_scan(std::forward<Args>(args)..., Partitioner());
    }
};

template<>
struct parallel_scan_wrapper<default_partitioner_tag>{
    template<typename... Args>
    void operator()(Args&&... args) {
    oneapi::tbb::parallel_scan(std::forward<Args>(args)...);
    }
};

// Test scan tag
//! \brief \ref interface
TEST_CASE("scan tags testing") {
    CHECK(oneapi::tbb::pre_scan_tag::is_final_scan()==false);
    CHECK(oneapi::tbb::final_scan_tag::is_final_scan()==true);
    CHECK((bool)oneapi::tbb::pre_scan_tag()==false);
    CHECK((bool)oneapi::tbb::final_scan_tag()==true);
}

//! Test parallel prefix sum calculation for body-based interface
//! \brief \ref requirement \ref interface
TEST_CASE_TEMPLATE("Test parallel scan with body", Partitioner, default_partitioner_tag, oneapi::tbb::simple_partitioner, oneapi::tbb::auto_partitioner) {
    std::vector<int> input(size);
    std::vector<int> output(size);
    std::vector<int> control(size);

    for(size_t i = 0; i < size; ++i) {
        input[i] = int(i / 2);
        if(i)
            control[i] = control[i-1] + input[i];
        else
            control[i] = input[i];
    }
    Body<int, std::plus<int>> body(input, output, 0);
    parallel_scan_wrapper<Partitioner>()(oneapi::tbb::blocked_range<std::size_t>(0U, size, 1U), body);
    CHECK((control == output));
}


//! Test parallel prefix sum calculation for scan-based interface
//! \brief \ref requirement \ref interface
TEST_CASE_TEMPLATE("Test parallel scan with body", Partitioner, default_partitioner_tag, oneapi::tbb::simple_partitioner, oneapi::tbb::auto_partitioner) {
    std::vector<std::size_t> input(size);
    std::vector<std::size_t> output(size);
    std::vector<std::size_t> control(size);

    for (std::size_t i = 0; i<size; ++i) {
        input[i] = i;
        if (i)
            control[i] = control[i-1]+input[i];
        else
            control[i] = input[i];
    }
    parallel_scan_wrapper<Partitioner>()(oneapi::tbb::blocked_range<std::size_t>(0U, size, 1U), std::size_t(0),
        [&](const oneapi::tbb::blocked_range<std::size_t>& r, std::size_t sum, bool is_final) -> std::size_t
        {
            std::size_t temp = sum;
            for (std::size_t i = r.begin(); i<r.end(); ++i) {
                temp = temp + input[i];
                if (is_final)
                    output[i] = temp;
            }
            return temp;
        },
        [](std::size_t a, std::size_t b) -> std::size_t
        {
            return a + b;
        });

    CHECK((control==output));
}

#if __TBB_CPP17_INVOKE_PRESENT
template <typename... Args>
void test_pscan_invoke(const std::vector<std::size_t>& desired_vector,
                       std::vector<std::size_t>& result_vector,
                       Args&&... args) {
    auto result = oneapi::tbb::parallel_scan(std::forward<Args>(args)...);
    CHECK(desired_vector == result_vector);
    CHECK(result.get() == result_vector.back());

    for (std::size_t& item : result_vector) item = 0;
}

//! Test that parallel_scan uses std::invoke to run the body
//! \brief \ref requirement
TEST_CASE("parallel_scan and std::invoke") {
    const std::size_t iterations = 1000000;
    std::vector<std::size_t> desired_vector(iterations);

    for (std::size_t i = 1; i < iterations; ++i) {
        desired_vector[i] = desired_vector[i - 1] + i;
    }

    std::vector<std::size_t> change_vector(iterations, 0);
    test_invoke::SmartRange<test_invoke::SmartValue> range(0, iterations, change_vector);
    test_invoke::SmartValue identity(0);

    auto scan = &test_invoke::SmartRange<test_invoke::SmartValue>::scan;
    auto combine = &test_invoke::SmartValue::operator+;

    test_pscan_invoke(desired_vector, change_vector, range, identity, scan, combine);
    test_pscan_invoke(desired_vector, change_vector, range, identity, scan, combine, oneapi::tbb::auto_partitioner());
    test_pscan_invoke(desired_vector, change_vector, range, identity, scan, combine, oneapi::tbb::simple_partitioner());
}

#endif // __TBB_CPP17_INVOKE_PRESENT