File: TestCuda_ReducerViewSizeLimit.cpp

package info (click to toggle)
kokkos 4.7.01-2
  • links: PTS, VCS
  • area: main
  • in suites: sid
  • size: 16,636 kB
  • sloc: cpp: 223,676; sh: 2,446; makefile: 2,437; python: 91; fortran: 4; ansic: 2
file content (159 lines) | stat: -rw-r--r-- 4,678 bytes parent folder | download | duplicates (2)
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