1 /*
2 * Copyright (c) 2002, 2011, Oracle and/or its affiliates. All rights reserved.
3 * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
4 *
5 * This code is free software; you can redistribute it and/or modify it
6 * under the terms of the GNU General Public License version 2 only, as
7 * published by the Free Software Foundation.
8 *
9 * This code is distributed in the hope that it will be useful, but WITHOUT
10 * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
11 * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
12 * version 2 for more details (a copy is included in the LICENSE file that
13 * accompanied this code).
14 *
15 * You should have received a copy of the GNU General Public License version
16 * 2 along with this work; if not, write to the Free Software Foundation,
17 * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA.
18 *
19 * Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA
20 * or visit www.oracle.com if you need additional information or have any
21 * questions.
22 *
23 */
24
25 #include "precompiled.hpp"
26 #include "gc_implementation/shared/gcUtil.hpp"
27
28 // Catch-all file for utility classes
29
30 float AdaptiveWeightedAverage::compute_adaptive_average(float new_sample,
31 float average) {
32 // We smooth the samples by not using weight() directly until we've
33 // had enough data to make it meaningful. We'd like the first weight
34 // used to be 1, the second to be 1/2, etc until we have
35 // OLD_THRESHOLD/weight samples.
36 unsigned count_weight = 0;
37
38 // Avoid division by zero if the counter wraps (7158457)
39 if (!is_old()) {
40 count_weight = OLD_THRESHOLD/count();
41 }
42
43 unsigned adaptive_weight = (MAX2(weight(), count_weight));
44
45 float new_avg = exp_avg(average, new_sample, adaptive_weight);
46
47 return new_avg;
48 }
49
50 void AdaptiveWeightedAverage::sample(float new_sample) {
51 increment_count();
52 assert(count() != 0,
53 "Wraparound -- history would be incorrectly discarded");
54
55 // Compute the new weighted average
56 float new_avg = compute_adaptive_average(new_sample, average());
57 set_average(new_avg);
58 _last_sample = new_sample;
59 }
60
61 void AdaptiveWeightedAverage::print() const {
62 print_on(tty);
63 }
64
65 void AdaptiveWeightedAverage::print_on(outputStream* st) const {
66 guarantee(false, "NYI");
67 }
68
69 void AdaptivePaddedAverage::print() const {
70 print_on(tty);
71 }
72
73 void AdaptivePaddedAverage::print_on(outputStream* st) const {
74 guarantee(false, "NYI");
75 }
76
77 void AdaptivePaddedNoZeroDevAverage::print() const {
78 print_on(tty);
79 }
80
81 void AdaptivePaddedNoZeroDevAverage::print_on(outputStream* st) const {
82 guarantee(false, "NYI");
83 }
84
85 void AdaptivePaddedAverage::sample(float new_sample) {
86 // Compute new adaptive weighted average based on new sample.
87 AdaptiveWeightedAverage::sample(new_sample);
88
89 // Now update the deviation and the padded average.
90 float new_avg = average();
91 float new_dev = compute_adaptive_average(fabsd(new_sample - new_avg),
92 deviation());
93 set_deviation(new_dev);
94 set_padded_average(new_avg + padding() * new_dev);
95 _last_sample = new_sample;
96 }
97
98 void AdaptivePaddedNoZeroDevAverage::sample(float new_sample) {
99 // Compute our parent classes sample information
100 AdaptiveWeightedAverage::sample(new_sample);
101
102 float new_avg = average();
103 if (new_sample != 0) {
104 // We only create a new deviation if the sample is non-zero
105 float new_dev = compute_adaptive_average(fabsd(new_sample - new_avg),
106 deviation());
107
108 set_deviation(new_dev);
109 }
110 set_padded_average(new_avg + padding() * deviation());
111 _last_sample = new_sample;
112 }
113
114 LinearLeastSquareFit::LinearLeastSquareFit(unsigned weight) :
115 _sum_x(0), _sum_x_squared(0), _sum_y(0), _sum_xy(0),
116 _intercept(0), _slope(0), _mean_x(weight), _mean_y(weight) {}
117
118 void LinearLeastSquareFit::update(double x, double y) {
119 _sum_x = _sum_x + x;
120 _sum_x_squared = _sum_x_squared + x * x;
121 _sum_y = _sum_y + y;
122 _sum_xy = _sum_xy + x * y;
123 _mean_x.sample(x);
124 _mean_y.sample(y);
125 assert(_mean_x.count() == _mean_y.count(), "Incorrect count");
126 if ( _mean_x.count() > 1 ) {
127 double slope_denominator;
128 slope_denominator = (_mean_x.count() * _sum_x_squared - _sum_x * _sum_x);
129 // Some tolerance should be injected here. A denominator that is
130 // nearly 0 should be avoided.
131
132 if (slope_denominator != 0.0) {
133 double slope_numerator;
134 slope_numerator = (_mean_x.count() * _sum_xy - _sum_x * _sum_y);
135 _slope = slope_numerator / slope_denominator;
136
137 // The _mean_y and _mean_x are decaying averages and can
138 // be used to discount earlier data. If they are used,
139 // first consider whether all the quantities should be
140 // kept as decaying averages.
141 // _intercept = _mean_y.average() - _slope * _mean_x.average();
142 _intercept = (_sum_y - _slope * _sum_x) / ((double) _mean_x.count());
143 }
144 }
145 }
146
147 double LinearLeastSquareFit::y(double x) {
148 double new_y;
149
150 if ( _mean_x.count() > 1 ) {
151 new_y = (_intercept + _slope * x);
152 return new_y;
153 } else {
154 return _mean_y.average();
155 }
156 }
157
158 // Both decrement_will_decrease() and increment_will_decrease() return
159 // true for a slope of 0. That is because a change is necessary before
160 // a slope can be calculated and a 0 slope will, in general, indicate
161 // that no calculation of the slope has yet been done. Returning true
162 // for a slope equal to 0 reflects the intuitive expectation of the
163 // dependence on the slope. Don't use the complement of these functions
164 // since that untuitive expectation is not built into the complement.
165 bool LinearLeastSquareFit::decrement_will_decrease() {
166 return (_slope >= 0.00);
167 }
168
169 bool LinearLeastSquareFit::increment_will_decrease() {
170 return (_slope <= 0.00);
171 }