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
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
|
// -*- C++ -*-
//==========================================================================
/**
* @file Stats.h
*
* $Id$
*
* @author David L. Levine
*/
//==========================================================================
#ifndef ACE_STATS_H
#define ACE_STATS_H
#include "ace/pre.h"
#include "ace/ACE_export.h"
#if !defined (ACE_LACKS_PRAGMA_ONCE)
# pragma once
#endif /* ACE_LACKS_PRAGMA_ONCE */
#include "ace/Unbounded_Queue.h"
#include "ace/Log_Msg.h"
#include "ace/Basic_Stats.h"
#include "ace/OS.h"
/**
* @class ACE_Stats_Value
*
* @brief Helper class for ACE_Stats.
*
* Container struct for 64-bit signed quantity and its
* precision. It would be nicer to use a fixed-point class, but
* this is sufficient. Users typically don't need to use this
* class directly; see ACE_Stats below.
*/
class ACE_Export ACE_Stats_Value
{
public:
/**
* Constructor, which requires precision in terms of number of
* decimal digits. The more variation in the data, and the greater
* the data values, the smaller the precision must be to avoid
* overflow in the standard deviation calculation. 3 might be a
* good value, or maybe 4. 5 will probably be too large for
* non-trivial data sets.
*/
ACE_Stats_Value (const u_int precision);
/// Accessor for precision.
u_int precision (void) const;
/// Set the whole_ field.
void whole (const ACE_UINT32);
/// Accessor for the whole_ field.
ACE_UINT32 whole (void) const;
/// Set the fractional_ field.
void fractional (const ACE_UINT32);
/// Accessor for the fractional_ field.
ACE_UINT32 fractional (void) const;
/// Calculates the maximum value of the fractional portion, given its
/// precision.
ACE_UINT32 fractional_field (void) const;
/**
* Access the value as an _unsigned_ 64 bit quantity. It scales the
* value up by <precision> decimal digits, so that no precision will
* be lost. It assumes that <whole_> is >= 0.
*/
void scaled_value (ACE_UINT64 &) const;
/// Print to stdout.
void dump (void) const;
private:
/// The integer portion of the value.
ACE_UINT32 whole_;
/// The fractional portion of the value.
ACE_UINT32 fractional_;
/**
* The number of decimal digits of precision represented by
* <fractional_>. Not declared const, so the only way to change it
* is via the assignment operator.
*/
u_int precision_;
ACE_UNIMPLEMENTED_FUNC (ACE_Stats_Value (void))
};
/**
* @class ACE_Stats
*
* @brief Provides simple statistical analysis.
*
* Simple statistical analysis package. Prominent features are:
* -# It does not use any floating point arithmetic.
* -# It handles positive and/or negative sample values. The
* sample value type is ACE_INT32.
* -# It uses 64 bit unsigned, but not 64 bit signed, quantities
* internally.
* -# It checks for overflow of internal state.
* -# It has no static variables of other than built-in types.
*
* Example usage:
*
* @verbatim
* ACE_Stats stats;
* for (u_int i = 0; i < n; ++i)
* {
* const ACE_UINT32 sample = ...;
* stats.sample (sample);
* }
* stats.print_summary (3);
* @endverbatim
*/
class ACE_Export ACE_Stats
{
public:
/// Default constructor.
ACE_Stats (void);
/// Provide a new sample. Returns 0 on success, -1 if it fails due
/// to running out of memory, or to rolling over of the sample count.
int sample (const ACE_INT32 value);
/// Access the number of samples provided so far.
ACE_UINT32 samples (void) const;
/// Value of the minimum sample provided so far.
ACE_INT32 min_value (void) const;
/// Value of the maximum sample provided so far.
ACE_INT32 max_value (void) const;
/**
* Access the mean of all samples provided so far. The fractional
* part is to the specified number of digits. E.g., 3 fractional
* digits specifies that the fractional part is in thousandths.
*/
void mean (ACE_Stats_Value &mean,
const ACE_UINT32 scale_factor = 1);
/// Access the standard deviation, whole and fractional parts. See
/// description of <mean> method for argument descriptions.
int std_dev (ACE_Stats_Value &std_dev,
const ACE_UINT32 scale_factor = 1);
/**
* Print summary statistics. If scale_factor is not 1, then the
* results are divided by it, i.e., each of the samples is scaled
* down by it. If internal overflow is reached with the specified
* scale factor, it successively tries to reduce it. Returns -1 if
* there is overflow even with a 0 scale factor.
*/
int print_summary (const u_int precision,
const ACE_UINT32 scale_factor = 1,
FILE * = stdout) const;
/// Initialize internal state.
void reset (void);
/// Utility division function, for ACE_UINT64 dividend.
static void quotient (const ACE_UINT64 dividend,
const ACE_UINT32 divisor,
ACE_Stats_Value "ient);
/// Utility division function, for ACE_Stats_Value dividend.
static void quotient (const ACE_Stats_Value ÷nd,
const ACE_UINT32 divisor,
ACE_Stats_Value "ient);
/**
* Sqrt function, which uses an oversimplified version of Newton's
* method. It's not fast, but it doesn't require floating point
* support.
*/
static void square_root (const ACE_UINT64 n,
ACE_Stats_Value &square_root);
/// Print summary statistics to stdout.
void dump (void) const;
private:
/// Internal indication of whether there has been overflow. Contains
/// the errno corresponding to the cause of overflow.
u_int overflow_;
/// Number of samples.
ACE_UINT32 number_of_samples_;
/// Minimum sample value.
ACE_INT32 min_;
/// Maximum sample value.
ACE_INT32 max_;
/// The samples.
ACE_Unbounded_Queue <ACE_INT32> samples_;
};
// ****************************************************************
/// A simple class to make throughput and latency analysis.
/**
*
* Keep the relevant information to perform throughput and latency
* analysis, including:
* -# Minimum, Average and Maximum latency
* -# Jitter for the latency
* -# Linear regression for throughput
* -# Accumulate results from several samples to obtain aggregated
* results, across several threads or experiments.
*
* @todo The idea behind this class was to use linear regression to
* determine if the throughput was linear or exhibited jitter.
* Unfortunately it never worked quite right, so only average
* throughput is computed.
*/
class ACE_Export ACE_Throughput_Stats : public ACE_Basic_Stats
{
public:
/// Constructor
ACE_Throughput_Stats (void);
/// Store one sample
void sample (ACE_UINT64 throughput, ACE_UINT64 latency);
/// Update the values to reflect the stats in @param throughput
void accumulate (const ACE_Throughput_Stats &throughput);
/// Print down the stats
void dump_results (const ACE_TCHAR* msg, ACE_UINT32 scale_factor);
/// Dump the average throughput stats.
static void dump_throughput (const ACE_TCHAR *msg,
ACE_UINT32 scale_factor,
ACE_UINT64 elapsed_time,
ACE_UINT32 samples_count);
private:
/// The last throughput measurement.
ACE_UINT64 throughput_last_;
#if 0
/// These are the fields that we should keep to perform linear
/// regression
//@{
///@}
ACE_UINT64 throughput_sum_x_;
ACE_UINT64 throughput_sum_x2_;
ACE_UINT64 throughput_sum_y_;
ACE_UINT64 throughput_sum_y2_;
ACE_UINT64 throughput_sum_xy_;
#endif /* 0 */
};
#if defined (__ACE_INLINE__)
# include "ace/Stats.i"
#endif /* __ACE_INLINE__ */
#include "ace/post.h"
#endif /* ! ACE_STATS_H */
|