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
|
/*
* Copyright 2008-2009 NVIDIA 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.
*/
#pragma once
#include <algorithm>
#include <numeric>
#include <vector>
#include <cusp/array1d.h>
#include <cusp/csr_matrix.h>
#include <thrust/count.h>
// TODO remove std::
namespace cusp
{
namespace detail
{
namespace host
{
namespace detail
{
template <typename Matrix>
size_t count_diagonals(const Matrix& csr, cusp::csr_format)
{
typedef typename Matrix::index_type IndexType;
cusp::array1d<bool,cusp::host_memory> occupied_diagonals(csr.num_rows + csr.num_cols, false);
for(size_t i = 0; i < csr.num_rows; i++)
{
for(IndexType jj = csr.row_offsets[i]; jj < csr.row_offsets[i+1]; jj++){
IndexType j = csr.column_indices[jj];
IndexType diagonal_offset = (csr.num_rows - i) + j; //offset shifted by + num_rows
occupied_diagonals[diagonal_offset] = true;
}
}
return thrust::count(occupied_diagonals.begin(), occupied_diagonals.end(), true);
}
template <typename Matrix>
size_t compute_max_entries_per_row(const Matrix& csr, cusp::csr_format)
{
size_t max_entries_per_row = 0;
for(size_t i = 0; i < csr.num_rows; i++)
max_entries_per_row = std::max<size_t>(max_entries_per_row, csr.row_offsets[i+1] - csr.row_offsets[i]);
return max_entries_per_row;
}
template <typename Matrix>
size_t compute_optimal_entries_per_row(const Matrix& csr,
float relative_speed,
size_t breakeven_threshold,
cusp::csr_format)
{
typedef typename Matrix::index_type IndexType;
// compute maximum row length
size_t max_cols_per_row = 0;
for(size_t i = 0; i < csr.num_rows; i++)
max_cols_per_row = std::max<size_t>(max_cols_per_row, csr.row_offsets[i+1] - csr.row_offsets[i]);
// compute distribution of nnz per row
std::vector<IndexType> histogram(max_cols_per_row + 1, 0);
for(size_t i = 0; i < csr.num_rows; i++)
histogram[csr.row_offsets[i+1] - csr.row_offsets[i]]++;
// compute optimal ELL column size
size_t num_cols_per_row = max_cols_per_row;
for(size_t i = 0, rows = csr.num_rows; i < max_cols_per_row; i++)
{
rows -= histogram[i]; //number of rows of length > i
if(relative_speed * rows < csr.num_rows || (size_t) rows < breakeven_threshold)
{
num_cols_per_row = i;
break;
}
}
return num_cols_per_row;
}
} // end namespace detail
template <typename Matrix>
size_t count_diagonals(const Matrix& m)
{
return cusp::detail::host::detail::count_diagonals(m, typename Matrix::format());
}
template <typename Matrix>
size_t compute_max_entries_per_row(const Matrix& m)
{
return cusp::detail::host::detail::compute_max_entries_per_row(m, typename Matrix::format());
}
////////////////////////////////////////////////////////////////////////////////
//! Compute Optimal Number of Columns per Row in the ELL part of the HYB format
//! Examines the distribution of nonzeros per row of the input CSR matrix to find
//! the optimal tradeoff between the ELL and COO portions of the hybrid (HYB)
//! sparse matrix format under the assumption that ELL performance is a fixed
//! multiple of COO performance. Furthermore, since ELL performance is also
//! sensitive to the absolute number of rows (and COO is not), a threshold is
//! used to ensure that the ELL portion contains enough rows to be worthwhile.
//! The default values were chosen empirically for a GTX280.
//!
//! @param csr CSR matrix
//! @param relative_speed Speed of ELL relative to COO (e.g. 2.0 -> ELL is twice as fast)
//! @param breakeven_threshold Minimum threshold at which ELL is faster than COO
////////////////////////////////////////////////////////////////////////////////
template <typename Matrix>
size_t compute_optimal_entries_per_row(const Matrix& m,
float relative_speed = 3.0f,
size_t breakeven_threshold = 4096)
{
return cusp::detail::host::detail::compute_optimal_entries_per_row
(m, relative_speed, breakeven_threshold, typename Matrix::format());
}
} // end namespace host
} // end namespace detail
} // end namespace cusp
|
