1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159
|
//@HEADER
// ************************************************************************
//
// Kokkos v. 4.0
// Copyright (2022) National Technology & Engineering
// Solutions of Sandia, LLC (NTESS).
//
// Under the terms of Contract DE-NA0003525 with NTESS,
// the U.S. Government retains certain rights in this software.
//
// Part of Kokkos, under the Apache License v2.0 with LLVM Exceptions.
// See https://kokkos.org/LICENSE for license information.
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
//
//@HEADER
#include <TestCuda_Category.hpp>
#include <Kokkos_Core.hpp>
namespace Test {
using ValueType = double;
using MemSpace = Kokkos::CudaSpace;
using Matrix2D = Kokkos::View<ValueType**, MemSpace>;
using Matrix3D = Kokkos::View<ValueType***, MemSpace>;
using Vector = Kokkos::View<ValueType*, MemSpace>;
namespace Impl {
struct ArrayReduceFunctor {
using value_type = ValueType[];
int value_count;
Matrix2D m;
ArrayReduceFunctor(const Matrix2D& m_) : value_count(m_.extent(1)), m(m_) {}
KOKKOS_INLINE_FUNCTION void operator()(const int i, value_type sum) const {
const int numVecs = value_count;
for (int j = 0; j < numVecs; ++j) {
sum[j] += m(i, j);
}
}
KOKKOS_INLINE_FUNCTION void init(value_type update) const {
const int numVecs = value_count;
for (int j = 0; j < numVecs; ++j) {
update[j] = 0.0;
}
}
KOKKOS_INLINE_FUNCTION void join(value_type update,
const value_type source) const {
const int numVecs = value_count;
for (int j = 0; j < numVecs; ++j) {
update[j] += source[j];
}
}
KOKKOS_INLINE_FUNCTION void final(value_type) const {}
};
struct MDArrayReduceFunctor {
using value_type = ValueType[];
int value_count;
Matrix3D m;
MDArrayReduceFunctor(const Matrix3D& m_) : value_count(m_.extent(2)), m(m_) {}
KOKKOS_INLINE_FUNCTION void operator()(const int i, const int j,
value_type sum) const {
const int numVecs = value_count;
for (int k = 0; k < numVecs; ++k) {
sum[k] += m(i, j, k);
}
}
KOKKOS_INLINE_FUNCTION void init(value_type update) const {
const int numVecs = value_count;
for (int j = 0; j < numVecs; ++j) {
update[j] = 0.0;
}
}
KOKKOS_INLINE_FUNCTION void final(value_type) const {}
};
struct ReduceViewSizeLimitTester {
const ValueType initValue = 3;
const size_t nGlobalEntries = 100;
const int testViewSize = 200;
const size_t expectedInitShmemLimit = 373584;
const unsigned initBlockSize = Kokkos::Impl::CudaTraits::WarpSize * 8;
void run_test_range() {
Matrix2D matrix;
Vector sum;
for (int i = 0; i < testViewSize; ++i) {
size_t sumInitShmemSize = (initBlockSize + 2) * sizeof(ValueType) * i;
Kokkos::resize(Kokkos::WithoutInitializing, sum, i);
Kokkos::resize(Kokkos::WithoutInitializing, matrix, nGlobalEntries, i);
Kokkos::deep_copy(matrix, initValue);
auto policy = Kokkos::RangePolicy<TEST_EXECSPACE>(0, nGlobalEntries);
auto functor = ArrayReduceFunctor(matrix);
if (sumInitShmemSize < expectedInitShmemLimit) {
EXPECT_NO_THROW(Kokkos::parallel_reduce(policy, functor, sum));
} else {
EXPECT_THROW(Kokkos::parallel_reduce(policy, functor, sum),
std::runtime_error);
}
}
}
void run_test_md_range_2D() {
Matrix3D matrix;
Vector sum;
for (int i = 0; i < testViewSize; ++i) {
size_t sumInitShmemSize = (initBlockSize + 2) * sizeof(ValueType) * i;
Kokkos::resize(Kokkos::WithoutInitializing, sum, i);
Kokkos::resize(Kokkos::WithoutInitializing, matrix, nGlobalEntries,
nGlobalEntries, i);
Kokkos::deep_copy(matrix, initValue);
auto policy = Kokkos::MDRangePolicy<Kokkos::Rank<2>>(
{0, 0}, {nGlobalEntries, nGlobalEntries});
auto functor = MDArrayReduceFunctor(matrix);
if (sumInitShmemSize < expectedInitShmemLimit) {
EXPECT_NO_THROW(Kokkos::parallel_reduce(policy, functor, sum));
} else {
EXPECT_THROW(Kokkos::parallel_reduce(policy, functor, sum),
std::runtime_error);
}
}
}
};
} // namespace Impl
TEST(cuda, reduceRangePolicyViewSizeLimit) {
Impl::ReduceViewSizeLimitTester reduceViewSizeLimitTester;
reduceViewSizeLimitTester.run_test_range();
}
TEST(cuda, reduceMDRangePolicyViewSizeLimit) {
Impl::ReduceViewSizeLimitTester reduceViewSizeLimitTester;
reduceViewSizeLimitTester.run_test_md_range_2D();
}
} // namespace Test
|