1 /* 2 3 Copyright (c) 2010-2018, Arvid Norberg 4 All rights reserved. 5 6 Redistribution and use in source and binary forms, with or without 7 modification, are permitted provided that the following conditions 8 are met: 9 10 * Redistributions of source code must retain the above copyright 11 notice, this list of conditions and the following disclaimer. 12 * Redistributions in binary form must reproduce the above copyright 13 notice, this list of conditions and the following disclaimer in 14 the documentation and/or other materials provided with the distribution. 15 * Neither the name of the author nor the names of its 16 contributors may be used to endorse or promote products derived 17 from this software without specific prior written permission. 18 19 THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" 20 AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 21 IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 22 ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE 23 LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR 24 CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF 25 SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS 26 INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN 27 CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) 28 ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE 29 POSSIBILITY OF SUCH DAMAGE. 30 31 */ 32 33 #ifndef TORRENT_SLIDING_AVERAGE_HPP_INCLUDED 34 #define TORRENT_SLIDING_AVERAGE_HPP_INCLUDED 35 36 #include <cstdint> 37 #include <cstdlib> // for std::abs 38 #include <limits> 39 #include <type_traits> // for is_integral 40 41 #include "libtorrent/assert.hpp" 42 43 namespace libtorrent { 44 45 // an exponential moving average accumulator. Add samples to it and it keeps 46 // track of a moving mean value and an average deviation 47 template <typename Int, Int inverted_gain> 48 struct sliding_average 49 { 50 static_assert(std::is_integral<Int>::value, "template argument must be integral"); 51 sliding_averagelibtorrent::sliding_average52 sliding_average(): m_mean(0), m_average_deviation(0), m_num_samples(0) {} 53 sliding_average(sliding_average const&) = default; 54 sliding_average& operator=(sliding_average const&) = default; 55 add_samplelibtorrent::sliding_average56 void add_sample(Int s) 57 { 58 TORRENT_ASSERT(s < std::numeric_limits<Int>::max() / 64); 59 // fixed point 60 s *= 64; 61 Int const deviation = (m_num_samples > 0) ? std::abs(m_mean - s) : 0; 62 63 if (m_num_samples < inverted_gain) 64 ++m_num_samples; 65 66 m_mean += (s - m_mean) / m_num_samples; 67 68 if (m_num_samples > 1) { 69 // the exact same thing for deviation off the mean except -1 on 70 // the samples, because the number of deviation samples always lags 71 // behind by 1 (you need to actual samples to have a single deviation 72 // sample). 73 m_average_deviation += (deviation - m_average_deviation) / (m_num_samples - 1); 74 } 75 } 76 meanlibtorrent::sliding_average77 Int mean() const { return m_num_samples > 0 ? (m_mean + 32) / 64 : 0; } avg_deviationlibtorrent::sliding_average78 Int avg_deviation() const { return m_num_samples > 1 ? (m_average_deviation + 32) / 64 : 0; } num_sampleslibtorrent::sliding_average79 int num_samples() const { return m_num_samples; } 80 81 private: 82 // both of these are fixed point values (* 64) 83 Int m_mean = 0; 84 Int m_average_deviation = 0; 85 // the number of samples we have received, but no more than inverted_gain 86 // this is the effective inverted_gain 87 int m_num_samples = 0; 88 }; 89 90 } 91 92 #endif 93