• file
• revisions
• annotate
• diff
• comparison
• raw

src/share/vm/gc_implementation/shared/gcUtil.hpp@e018e6884bd8 (annotated)

src/share/vm/gc_implementation/shared/gcUtil.hpp

Wed, 23 Dec 2009 09:23:54 -0800

author

ysr

date

Wed, 23 Dec 2009 09:23:54 -0800

changeset 1580

e018e6884bd8

parent 772

9ee9cf798b59

child 1907

c18cbe5936b8

permissions

-rw-r--r--

6631166: CMS: better heuristics when combatting fragmentation
Summary: Autonomic per-worker free block cache sizing, tunable coalition policies, fixes to per-size block statistics, retuned gain and bandwidth of some feedback loop filters to allow quicker reactivity to abrupt changes in ambient demand, and other heuristics to reduce fragmentation of the CMS old gen. Also tightened some assertions, including those related to locking.
Reviewed-by: jmasa

duke@435	1	/*
xdono@772	2	* Copyright 2002-2008 Sun Microsystems, Inc. All Rights Reserved.
duke@435	3	* DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
duke@435	4	*
duke@435	5	* This code is free software; you can redistribute it and/or modify it
duke@435	6	* under the terms of the GNU General Public License version 2 only, as
duke@435	7	* published by the Free Software Foundation.
duke@435	8	*
duke@435	9	* This code is distributed in the hope that it will be useful, but WITHOUT
duke@435	10	* ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
duke@435	11	* FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
duke@435	12	* version 2 for more details (a copy is included in the LICENSE file that
duke@435	13	* accompanied this code).
duke@435	14	*
duke@435	15	* You should have received a copy of the GNU General Public License version
duke@435	16	* 2 along with this work; if not, write to the Free Software Foundation,
duke@435	17	* Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA.
duke@435	18	*
duke@435	19	* Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
duke@435	20	* CA 95054 USA or visit www.sun.com if you need additional information or
duke@435	21	* have any questions.
duke@435	22	*
duke@435	23	*/
duke@435	24
duke@435	25	// Catch-all file for utility classes
duke@435	26
duke@435	27	// A weighted average maintains a running, weighted average
duke@435	28	// of some float value (templates would be handy here if we
duke@435	29	// need different types).
duke@435	30	//
duke@435	31	// The average is adaptive in that we smooth it for the
duke@435	32	// initial samples; we don't use the weight until we have
duke@435	33	// enough samples for it to be meaningful.
duke@435	34	//
duke@435	35	// This serves as our best estimate of a future unknown.
duke@435	36	//
duke@435	37	class AdaptiveWeightedAverage : public CHeapObj {
duke@435	38	private:
duke@435	39	float _average; // The last computed average
duke@435	40	unsigned _sample_count; // How often we've sampled this average
duke@435	41	unsigned _weight; // The weight used to smooth the averages
duke@435	42	// A higher weight favors the most
duke@435	43	// recent data.
duke@435	44
duke@435	45	protected:
duke@435	46	float _last_sample; // The last value sampled.
duke@435	47
duke@435	48	void increment_count() { _sample_count++; }
duke@435	49	void set_average(float avg) { _average = avg; }
duke@435	50
duke@435	51	// Helper function, computes an adaptive weighted average
duke@435	52	// given a sample and the last average
duke@435	53	float compute_adaptive_average(float new_sample, float average);
duke@435	54
duke@435	55	public:
duke@435	56	// Input weight must be between 0 and 100
ysr@1580	57	AdaptiveWeightedAverage(unsigned weight, float avg = 0.0) :
ysr@1580	58	_average(avg), _sample_count(0), _weight(weight), _last_sample(0.0) {
duke@435	59	}
duke@435	60
iveresov@703	61	void clear() {
iveresov@703	62	_average = 0;
iveresov@703	63	_sample_count = 0;
iveresov@703	64	_last_sample = 0;
iveresov@703	65	}
iveresov@703	66
ysr@1580	67	// Useful for modifying static structures after startup.
ysr@1580	68	void modify(size_t avg, unsigned wt, bool force = false) {
ysr@1580	69	assert(force, "Are you sure you want to call this?");
ysr@1580	70	_average = (float)avg;
ysr@1580	71	_weight = wt;
ysr@1580	72	}
ysr@1580	73
duke@435	74	// Accessors
duke@435	75	float average() const { return _average; }
duke@435	76	unsigned weight() const { return _weight; }
duke@435	77	unsigned count() const { return _sample_count; }
duke@435	78	float last_sample() const { return _last_sample; }
duke@435	79
duke@435	80	// Update data with a new sample.
duke@435	81	void sample(float new_sample);
duke@435	82
duke@435	83	static inline float exp_avg(float avg, float sample,
duke@435	84	unsigned int weight) {
duke@435	85	assert(0 <= weight && weight <= 100, "weight must be a percent");
duke@435	86	return (100.0F - weight) * avg / 100.0F + weight * sample / 100.0F;
duke@435	87	}
duke@435	88	static inline size_t exp_avg(size_t avg, size_t sample,
duke@435	89	unsigned int weight) {
duke@435	90	// Convert to float and back to avoid integer overflow.
duke@435	91	return (size_t)exp_avg((float)avg, (float)sample, weight);
duke@435	92	}
ysr@1580	93
ysr@1580	94	// Printing
ysr@1580	95	void print_on(outputStream* st) const;
ysr@1580	96	void print() const;
duke@435	97	};
duke@435	98
duke@435	99
duke@435	100	// A weighted average that includes a deviation from the average,
duke@435	101	// some multiple of which is added to the average.
duke@435	102	//
duke@435	103	// This serves as our best estimate of an upper bound on a future
duke@435	104	// unknown.
duke@435	105	class AdaptivePaddedAverage : public AdaptiveWeightedAverage {
duke@435	106	private:
duke@435	107	float _padded_avg; // The last computed padded average
duke@435	108	float _deviation; // Running deviation from the average
duke@435	109	unsigned _padding; // A multiple which, added to the average,
duke@435	110	// gives us an upper bound guess.
duke@435	111
duke@435	112	protected:
duke@435	113	void set_padded_average(float avg) { _padded_avg = avg; }
duke@435	114	void set_deviation(float dev) { _deviation = dev; }
duke@435	115
duke@435	116	public:
duke@435	117	AdaptivePaddedAverage() :
duke@435	118	AdaptiveWeightedAverage(0),
duke@435	119	_padded_avg(0.0), _deviation(0.0), _padding(0) {}
duke@435	120
duke@435	121	AdaptivePaddedAverage(unsigned weight, unsigned padding) :
duke@435	122	AdaptiveWeightedAverage(weight),
duke@435	123	_padded_avg(0.0), _deviation(0.0), _padding(padding) {}
duke@435	124
duke@435	125	// Placement support
duke@435	126	void* operator new(size_t ignored, void* p) { return p; }
duke@435	127	// Allocator
duke@435	128	void* operator new(size_t size) { return CHeapObj::operator new(size); }
duke@435	129
duke@435	130	// Accessor
duke@435	131	float padded_average() const { return _padded_avg; }
duke@435	132	float deviation() const { return _deviation; }
duke@435	133	unsigned padding() const { return _padding; }
duke@435	134
iveresov@703	135	void clear() {
iveresov@703	136	AdaptiveWeightedAverage::clear();
iveresov@703	137	_padded_avg = 0;
iveresov@703	138	_deviation = 0;
iveresov@703	139	}
iveresov@703	140
duke@435	141	// Override
duke@435	142	void sample(float new_sample);
ysr@1580	143
ysr@1580	144	// Printing
ysr@1580	145	void print_on(outputStream* st) const;
ysr@1580	146	void print() const;
duke@435	147	};
duke@435	148
duke@435	149	// A weighted average that includes a deviation from the average,
duke@435	150	// some multiple of which is added to the average.
duke@435	151	//
duke@435	152	// This serves as our best estimate of an upper bound on a future
duke@435	153	// unknown.
duke@435	154	// A special sort of padded average: it doesn't update deviations
duke@435	155	// if the sample is zero. The average is allowed to change. We're
duke@435	156	// preventing the zero samples from drastically changing our padded
duke@435	157	// average.
duke@435	158	class AdaptivePaddedNoZeroDevAverage : public AdaptivePaddedAverage {
duke@435	159	public:
duke@435	160	AdaptivePaddedNoZeroDevAverage(unsigned weight, unsigned padding) :
duke@435	161	AdaptivePaddedAverage(weight, padding) {}
duke@435	162	// Override
duke@435	163	void sample(float new_sample);
ysr@1580	164
ysr@1580	165	// Printing
ysr@1580	166	void print_on(outputStream* st) const;
ysr@1580	167	void print() const;
duke@435	168	};
ysr@1580	169
duke@435	170	// Use a least squares fit to a set of data to generate a linear
duke@435	171	// equation.
duke@435	172	// y = intercept + slope * x
duke@435	173
duke@435	174	class LinearLeastSquareFit : public CHeapObj {
duke@435	175	double _sum_x; // sum of all independent data points x
duke@435	176	double _sum_x_squared; // sum of all independent data points x**2
duke@435	177	double _sum_y; // sum of all dependent data points y
duke@435	178	double _sum_xy; // sum of all x * y.
duke@435	179	double _intercept; // constant term
duke@435	180	double _slope; // slope
duke@435	181	// The weighted averages are not currently used but perhaps should
duke@435	182	// be used to get decaying averages.
duke@435	183	AdaptiveWeightedAverage _mean_x; // weighted mean of independent variable
duke@435	184	AdaptiveWeightedAverage _mean_y; // weighted mean of dependent variable
duke@435	185
duke@435	186	public:
duke@435	187	LinearLeastSquareFit(unsigned weight);
duke@435	188	void update(double x, double y);
duke@435	189	double y(double x);
duke@435	190	double slope() { return _slope; }
duke@435	191	// Methods to decide if a change in the dependent variable will
duke@435	192	// achive a desired goal. Note that these methods are not
duke@435	193	// complementary and both are needed.
duke@435	194	bool decrement_will_decrease();
duke@435	195	bool increment_will_decrease();
duke@435	196	};
duke@435	197
duke@435	198	class GCPauseTimer : StackObj {
duke@435	199	elapsedTimer* _timer;
duke@435	200	public:
duke@435	201	GCPauseTimer(elapsedTimer* timer) {
duke@435	202	_timer = timer;
duke@435	203	_timer->stop();
duke@435	204	}
duke@435	205	~GCPauseTimer() {
duke@435	206	_timer->start();
duke@435	207	}
duke@435	208	};