Sat, 27 Sep 2008 00:33:13 -0700
6740923: NUMA allocator: Ensure the progress of adaptive chunk resizing
Summary: Treat a chuck where the allocation has failed as fully used.
Reviewed-by: ysr
1 /*
2 * Copyright 2002-2005 Sun Microsystems, Inc. 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 Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
20 * CA 95054 USA or visit www.sun.com if you need additional information or
21 * have any questions.
22 *
23 */
25 // Catch-all file for utility classes
27 // A weighted average maintains a running, weighted average
28 // of some float value (templates would be handy here if we
29 // need different types).
30 //
31 // The average is adaptive in that we smooth it for the
32 // initial samples; we don't use the weight until we have
33 // enough samples for it to be meaningful.
34 //
35 // This serves as our best estimate of a future unknown.
36 //
37 class AdaptiveWeightedAverage : public CHeapObj {
38 private:
39 float _average; // The last computed average
40 unsigned _sample_count; // How often we've sampled this average
41 unsigned _weight; // The weight used to smooth the averages
42 // A higher weight favors the most
43 // recent data.
45 protected:
46 float _last_sample; // The last value sampled.
48 void increment_count() { _sample_count++; }
49 void set_average(float avg) { _average = avg; }
51 // Helper function, computes an adaptive weighted average
52 // given a sample and the last average
53 float compute_adaptive_average(float new_sample, float average);
55 public:
56 // Input weight must be between 0 and 100
57 AdaptiveWeightedAverage(unsigned weight) :
58 _average(0.0), _sample_count(0), _weight(weight), _last_sample(0.0) {
59 }
61 void clear() {
62 _average = 0;
63 _sample_count = 0;
64 _last_sample = 0;
65 }
67 // Accessors
68 float average() const { return _average; }
69 unsigned weight() const { return _weight; }
70 unsigned count() const { return _sample_count; }
71 float last_sample() const { return _last_sample; }
73 // Update data with a new sample.
74 void sample(float new_sample);
76 static inline float exp_avg(float avg, float sample,
77 unsigned int weight) {
78 assert(0 <= weight && weight <= 100, "weight must be a percent");
79 return (100.0F - weight) * avg / 100.0F + weight * sample / 100.0F;
80 }
81 static inline size_t exp_avg(size_t avg, size_t sample,
82 unsigned int weight) {
83 // Convert to float and back to avoid integer overflow.
84 return (size_t)exp_avg((float)avg, (float)sample, weight);
85 }
86 };
89 // A weighted average that includes a deviation from the average,
90 // some multiple of which is added to the average.
91 //
92 // This serves as our best estimate of an upper bound on a future
93 // unknown.
94 class AdaptivePaddedAverage : public AdaptiveWeightedAverage {
95 private:
96 float _padded_avg; // The last computed padded average
97 float _deviation; // Running deviation from the average
98 unsigned _padding; // A multiple which, added to the average,
99 // gives us an upper bound guess.
101 protected:
102 void set_padded_average(float avg) { _padded_avg = avg; }
103 void set_deviation(float dev) { _deviation = dev; }
105 public:
106 AdaptivePaddedAverage() :
107 AdaptiveWeightedAverage(0),
108 _padded_avg(0.0), _deviation(0.0), _padding(0) {}
110 AdaptivePaddedAverage(unsigned weight, unsigned padding) :
111 AdaptiveWeightedAverage(weight),
112 _padded_avg(0.0), _deviation(0.0), _padding(padding) {}
114 // Placement support
115 void* operator new(size_t ignored, void* p) { return p; }
116 // Allocator
117 void* operator new(size_t size) { return CHeapObj::operator new(size); }
119 // Accessor
120 float padded_average() const { return _padded_avg; }
121 float deviation() const { return _deviation; }
122 unsigned padding() const { return _padding; }
124 void clear() {
125 AdaptiveWeightedAverage::clear();
126 _padded_avg = 0;
127 _deviation = 0;
128 }
130 // Override
131 void sample(float new_sample);
132 };
134 // A weighted average that includes a deviation from the average,
135 // some multiple of which is added to the average.
136 //
137 // This serves as our best estimate of an upper bound on a future
138 // unknown.
139 // A special sort of padded average: it doesn't update deviations
140 // if the sample is zero. The average is allowed to change. We're
141 // preventing the zero samples from drastically changing our padded
142 // average.
143 class AdaptivePaddedNoZeroDevAverage : public AdaptivePaddedAverage {
144 public:
145 AdaptivePaddedNoZeroDevAverage(unsigned weight, unsigned padding) :
146 AdaptivePaddedAverage(weight, padding) {}
147 // Override
148 void sample(float new_sample);
149 };
150 // Use a least squares fit to a set of data to generate a linear
151 // equation.
152 // y = intercept + slope * x
154 class LinearLeastSquareFit : public CHeapObj {
155 double _sum_x; // sum of all independent data points x
156 double _sum_x_squared; // sum of all independent data points x**2
157 double _sum_y; // sum of all dependent data points y
158 double _sum_xy; // sum of all x * y.
159 double _intercept; // constant term
160 double _slope; // slope
161 // The weighted averages are not currently used but perhaps should
162 // be used to get decaying averages.
163 AdaptiveWeightedAverage _mean_x; // weighted mean of independent variable
164 AdaptiveWeightedAverage _mean_y; // weighted mean of dependent variable
166 public:
167 LinearLeastSquareFit(unsigned weight);
168 void update(double x, double y);
169 double y(double x);
170 double slope() { return _slope; }
171 // Methods to decide if a change in the dependent variable will
172 // achive a desired goal. Note that these methods are not
173 // complementary and both are needed.
174 bool decrement_will_decrease();
175 bool increment_will_decrease();
176 };
178 class GCPauseTimer : StackObj {
179 elapsedTimer* _timer;
180 public:
181 GCPauseTimer(elapsedTimer* timer) {
182 _timer = timer;
183 _timer->stop();
184 }
185 ~GCPauseTimer() {
186 _timer->start();
187 }
188 };