Commit | Line | Data |
e98cedbf |
1 | package Devel::Size; |
2 | |
e98cedbf |
3 | use strict; |
0430b7f7 |
4 | use vars qw($VERSION @ISA @EXPORT @EXPORT_OK %EXPORT_TAGS $warn); |
e98cedbf |
5 | |
6ea94d90 |
6 | require 5.006; |
e98cedbf |
7 | require Exporter; |
8 | require DynaLoader; |
9 | |
a6ea0805 |
10 | @ISA = qw(Exporter DynaLoader); |
e98cedbf |
11 | |
e98cedbf |
12 | # This allows declaration use Devel::Size ':all'; |
a6ea0805 |
13 | %EXPORT_TAGS = ( 'all' => [ qw( |
0bff12d8 |
14 | size total_size |
e98cedbf |
15 | ) ] ); |
16 | |
a6ea0805 |
17 | @EXPORT_OK = ( @{ $EXPORT_TAGS{'all'} } ); |
e98cedbf |
18 | |
0430b7f7 |
19 | @EXPORT = qw( ); |
8b959707 |
20 | $VERSION = '0.70'; |
e98cedbf |
21 | |
22 | bootstrap Devel::Size $VERSION; |
23 | |
78dfb4e7 |
24 | $warn = 1; |
ebb2c5b9 |
25 | |
e98cedbf |
26 | 1; |
27 | __END__ |
e98cedbf |
28 | |
6ea94d90 |
29 | =pod |
30 | |
e98cedbf |
31 | =head1 NAME |
32 | |
0bff12d8 |
33 | Devel::Size - Perl extension for finding the memory usage of Perl variables |
e98cedbf |
34 | |
35 | =head1 SYNOPSIS |
36 | |
0bff12d8 |
37 | use Devel::Size qw(size total_size); |
e98cedbf |
38 | |
0bff12d8 |
39 | my $size = size("A string"); |
40 | |
41 | my @foo = (1, 2, 3, 4, 5); |
42 | my $other_size = size(\@foo); |
43 | |
44 | my $foo = {a => [1, 2, 3], |
5c2e1b12 |
45 | b => {a => [1, 3, 4]} |
46 | }; |
5a83b7cf |
47 | my $total_size = total_size($foo); |
5c2e1b12 |
48 | |
e98cedbf |
49 | =head1 DESCRIPTION |
50 | |
5a83b7cf |
51 | This module figures out the real size of Perl variables in bytes, as |
52 | accurately as possible. |
53 | |
0bff12d8 |
54 | Call functions with a reference to the variable you want the size |
55 | of. If the variable is a plain scalar it returns the size of |
5a83b7cf |
56 | this scalar. If the variable is a hash or an array, use a reference |
0bff12d8 |
57 | when calling. |
58 | |
59 | =head1 FUNCTIONS |
60 | |
61 | =head2 size($ref) |
e98cedbf |
62 | |
5c2e1b12 |
63 | The C<size> function returns the amount of memory the variable |
0bff12d8 |
64 | returns. If the variable is a hash or an array, it only reports |
65 | the amount used by the structure, I<not> the contents. |
66 | |
67 | =head2 total_size($ref) |
5c2e1b12 |
68 | |
0bff12d8 |
69 | The C<total_size> function will traverse the variable and look |
70 | at the sizes of contents. Any references contained in the variable |
71 | will also be followed, so this function can be used to get the |
72 | total size of a multidimensional data structure. At the moment |
73 | there is no way to get the size of an array or a hash and its |
74 | elements without using this function. |
5c2e1b12 |
75 | |
b98fcdb9 |
76 | =head1 EXPORT |
e98cedbf |
77 | |
0bff12d8 |
78 | None but default, but optionally C<size> and C<total_size>. |
e98cedbf |
79 | |
b98fcdb9 |
80 | =head1 UNDERSTANDING MEMORY ALLOCATION |
81 | |
82 | Please note that the following discussion of memory allocation in perl |
83 | is based on the perl 5.8.0 sources. While this is generally |
84 | applicable to all versions of perl, some of the gory details are |
85 | omitted. It also makes some presumptions on how your system memory |
86 | allocator works so, while it will be generally correct, it may not |
87 | exactly reflect your system. (Generally the only issue is the size of |
88 | the constant values we'll talk about, not their existence) |
89 | |
90 | =head2 The C library |
91 | |
92 | It's important firtst to understand how your OS and libraries handle |
93 | memory. When the perl interpreter needs some memory, it asks the C |
94 | runtime library for it, using the C<malloc()> call. C<malloc> has one |
95 | parameter, the size of the memory allocation you want, and returns a |
96 | pointer to that memory. C<malloc> also makes sure that the pointer it |
97 | returns to you is properly aligned. When you're done with the memory |
98 | you hand it back to the library with the C<free()> call. C<free> has |
99 | one parameter, the pointer that C<malloc> returned. There are a couple of interesting ramifications to this. |
100 | |
101 | Because malloc has to return an aligned pointer, it will round up the |
102 | memory allocation to make sure that the memory it returns is aligned |
103 | right. What that alignment is depends on your CPU, OS, and compiler |
104 | settings, but things are generally aligned to either a 4 or 8 byte |
105 | boundary. That means that if you ask for 1 byte, C<malloc> will |
106 | silently round up to either 4 or 8 bytes, though it doesn't tell the |
107 | program making the request, so the extra memory can't be used. |
108 | |
109 | Since C<free> isn't given the size of the memory chunk you're |
110 | freeing, it has to track it another way. Most libraries do this by |
111 | tacking on a length field just before the memory it hands to your |
112 | program. (It's put before the beginning rather than after the end |
113 | because it's less likely to get mangled by program bugs) This size |
114 | field is the size of your platform integer, Generally either 4 or 8 |
115 | bytes. |
116 | |
117 | So, if you asked for 1 byte, malloc would build something like this: |
118 | |
119 | +------------------+ |
120 | | 4 byte length | |
121 | +------------------+ <----- the pointer malloc returns |
122 | | your 1 byte | |
123 | +------------------+ |
124 | | 3 bytes padding | |
125 | +------------------+ |
126 | |
127 | As you can see, you asked for 1 byte but C<malloc> used 8. If your |
128 | integers were 8 bytes rather than 4, C<malloc> would have used 16 bytes |
129 | to satisfy your 1 byte request. |
130 | |
131 | The C memory allocation system also keeps a list of free memory |
132 | chunks, so it can recycle freed memory. For performance reasons, some |
133 | C memory allocation systems put a limit to the number of free |
134 | segments that are on the free list, or only search through a small |
135 | number of memory chunks waiting to be recycled before just |
136 | allocating more memory from the system. |
137 | |
138 | The memory allocation system tries to keep as few chunks on the free |
139 | list as possible. It does this by trying to notice if there are two |
140 | adjacent chunks of memory on the free list and, if there are, |
141 | coalescing them into a single larger chunk. This works pretty well, |
142 | but there are ways to have a lot of memory on the free list yet still |
143 | not have anything that can be allocated. If a program allocates one |
144 | million eight-byte chunks, for example, then frees every other chunk, |
145 | there will be four million bytes of memory on the free list, but none |
146 | of that memory can be handed out to satisfy a request for 10 |
147 | bytes. This is what's referred to as a fragmented free list, and can |
148 | be one reason why your program could have a lot of free memory yet |
149 | still not be able to allocate more, or have a huge process size and |
150 | still have almost no memory actually allocated to the program running. |
151 | |
152 | =head2 Perl |
153 | |
154 | Perl's memory allocation scheme is a bit convoluted, and more complex |
0430b7f7 |
155 | than can really be addressed here, but there is one common spot where Perl's |
b98fcdb9 |
156 | memory allocation is unintuitive, and that's for hash keys. |
157 | |
158 | When you have a hash, each entry has a structure that points to the |
159 | key and the value for that entry. The value is just a pointer to the |
160 | scalar in the entry, and doesn't take up any special amount of |
161 | memory. The key structure holds the hash value for the key, the key |
162 | length, and the key string. (The entry and key structures are |
163 | separate so perl can potentially share keys across multiple hashes) |
164 | |
165 | The entry structure has three pointers in it, and takes up either 12 |
166 | or 24 bytes, depending on whether you're on a 32 bit or 64 bit |
167 | system. Since these structures are of fixed size, perl can keep a big |
168 | pool of them internally (generally called an arena) so it doesn't |
169 | have to allocate memory for each one. |
170 | |
171 | The key structure, though, is of variable length because the key |
172 | string is of variable length, so perl has to ask the system for a |
173 | memory allocation for each key. The base size of this structure is |
174 | 8 or 16 bytes (once again, depending on whether you're on a 32 bit or |
175 | 64 bit system) plus the string length plus two bytes. |
176 | |
177 | Since this memory has to be allocated from the system there's the |
178 | malloc size-field overhead (4 or 8 bytes) plus the alignment bytes (0 |
179 | to 7, depending on your system and the key length) |
180 | that get added on to the chunk perl requests. If the key is only 1 |
181 | character, and you're on a 32 bit system, the allocation will be 16 |
182 | bytes. If the key is 7 characters then the allocation is 24 bytes on |
183 | a 32 bit system. If you're on a 64 bit system the numbers get even |
184 | larger. |
185 | |
186 | This does mean that hashes eat up a I<lot> of memory, both in memory |
187 | Devel::Size can track (the memory actually in the structures and |
188 | strings) and that it can't (the malloc alignment and length overhead). |
189 | |
190 | =head1 DANGERS |
191 | |
192 | Devel::Size, because of the way it works, can consume a |
193 | considerable amount of memory as it runs. It will use five |
194 | pointers, two integers, and two bytes worth of storage, plus |
195 | potential alignment and bucket overhead, per thing it looks at. This |
196 | memory is released at the end, but it may fragment your free pool, |
197 | and will definitely expand your process' memory footprint. |
198 | |
5073b933 |
199 | =head1 Messages: texts originating from this module. |
200 | |
201 | =head2 Errors |
202 | |
203 | =over 4 |
204 | |
205 | =item "Devel::Size: Unknown variable type" |
206 | |
207 | The thing (or something contained within it) that you gave to |
208 | total_size() was unrecognisable as a Perl entity. |
209 | |
210 | =back |
211 | |
212 | =head2 warnings |
213 | |
214 | These messages warn you that for some types, the sizes calculated may not include |
215 | everything that could be associated with those types. The differences are usually |
216 | insignificant for most uses of this module. |
217 | |
218 | These may be disabled by setting |
219 | |
220 | $Devel::Size::warn = 0 |
221 | |
222 | =over 4 |
223 | |
224 | =item "Devel::Size: Calculated sizes for CVs are incomplete" |
225 | |
226 | =item "Devel::Size: Calculated sizes for FMs are incomplete" |
227 | |
5073b933 |
228 | =back |
229 | |
e98cedbf |
230 | =head1 BUGS |
231 | |
fea63ffa |
232 | Doesn't currently walk all the bits for code refs, formats, and |
6a9ad7ec |
233 | IO. Those throw a warning, but a minimum size for them is returned. |
e98cedbf |
234 | |
b98fcdb9 |
235 | Devel::Size only counts the memory that perl actually allocates. It |
236 | doesn't count 'dark' memory--memory that is lost due to fragmented free lists, |
237 | allocation alignments, or C library overhead. |
238 | |
e98cedbf |
239 | =head1 AUTHOR |
240 | |
241 | Dan Sugalski dan@sidhe.org |
242 | |
98ecbbc6 |
243 | Small portion taken from the B module as shipped with perl 5.6.2. |
244 | |
0430b7f7 |
245 | Maintained now by Tels <http://bloodgate.com> |
246 | |
98ecbbc6 |
247 | =head1 COPYRIGHT |
248 | |
6ea94d90 |
249 | Copyright (C) 2005 Dan Sugalski, Copyright (C) 2007-2008 Tels |
98ecbbc6 |
250 | |
251 | This module is free software; you can redistribute it and/or modify it |
5a83b7cf |
252 | under the same terms as Perl v5.8.8. |
98ecbbc6 |
253 | |
e98cedbf |
254 | =head1 SEE ALSO |
255 | |
0430b7f7 |
256 | perl(1), L<Devel::Size::Report>. |
e98cedbf |
257 | |
258 | =cut |