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
|
0.19 20 September 2010
Checking tool chain with perl.5.13.5.
0.18 3 October 2007
Hopefully fixed test-suite for none thread enabled Perls. Updated
to use new VERSION feature of Devel::Required.
0.17 7 January 2004
Fixed some documentation inconsistencies. Added documentation on what
is different from Thread::Queue::Any.
Made default optimization dependent on the version of Perl used.
Added test for it.
Removed mention of Storable: replaced it with Thread::Serialize.
Added Thread::Serialize as a dependency, just for documentation sake
really.
0.16 28 December 2003
Added automatic required modules update using Devel::Required.
Added dependency to load.pm.
0.15 9 November 2003
A problem in (threaded) 5.8.2 made Thread::Conveyor segfault in the
test-suite. The problem was fortunately easy to circumvent: just don't
use subroutines consisting of {} as an object method (in the case of
Thread::Conveyor, the shutdown() method).
0.14 11 August 2003
Cleaned up Makefile.PL and updated copyright info. Verified everything
works with 5.8.1. Updated dependency of Thread::Tie.
0.13 30 September 2002
Use "load.pm" instead of AutoLoader. Changed dependency on Thread::Tie
to version 0.08 (which also uses load.pm).
Added check for availability of methods again: load.pm does support
that, whereas AutoLoader does not.
0.12 27 September 2002
Ask found another problem in T::C::Tied::clean_dontwait. Fixed.
Also left warnings enabled in the test-suite last time, which caused
warnings in the make test. Changed code so that the warnings are not
emitted anymore.
0.11 26 September 2002
Fixed some warnings in T::C::Throttled, spotted by Ask Bjoern Hansen.
Removed "our" from $VERSION and @ISA, should shave off some bytes in
memory usage, as found from testing with Benchmark::Thread::Size.
0.10 1 September 2002
Replaced dependency on Thread::Serialize by dependency on Thread::Tie
as the default, memory optimized implementation, now uses Thread::Tie.
Changed Thread::Conveyor so that it uses T::C::Tied for unthrottled
belts when optimizing for memory.
Changed T::C::Throttled so that it makes its own decision on which
belt implementation to be used, either T::C::Array (optimized for
CPU) or T::C::Tied (optimized for memory).
Added "semaphore" object method to T::C::Array for obtaining the
lock() semaphore, needed by T::C::Throttled.
Removed T::C::Thread from the distribution. This shaves about 1200
bytes off of the distribution package.
Replaced Thread::Conveyor::Thread by Thread::Conveyor::Tied, which
uses Thread::Tie as its shared array implementation. This has the
side-effect to being almost twice as slow but saving 18% of
memory (at least in the test-suite). So there is now a true CPU
versus memory trade-off. Main difference in CPU is caused by the
fact that whole array fetches aren't possible using the tie()
interface (this is a loop with FETCH()), wherease T::C::Thread was
optimized to deal with that case). Main difference in memory is
caused by the fact that only one thread (the Thread::Tie default
thread) is used for all belts, whereas before each belt had its own
thread. So the savings in memory could in real world situations, be
quite a lot more than the 18% seen in the test-suite.
0.09 30 August 2002
Changed all modules to use AutoLoader to defer loading of necessary
subroutines to when they are actually needed. This should save
memory and CPU for larger programs, or with large numbers of threads.
The test-suite only marginally takes more memory at the same CPU
usage: overhead of compiling is levelled out with the overhead of
cloning pre-compiled routines.
0.08 20 August 2002
Dave Mitchell showed me the error of my ways with regards to client
server thread programming. Fixed the handler optimized for memory so
that no yield()s are necessary anymore. The whole thing now gets
through the test suite more than 25% faster for that part.
Scalability to many more threads should now be a lot better too.
0.07 14 August 2002
Removed (possibly premature) optimization from T::C::Thread
_handler: it now doesn't zap the local threadspace anymore.
0.06 13 August 2002
Handed over freezing and thawing operations to Thread::Serialize.
Added dependency to Thread::Serialize in Makefile.PL.
0.05 6 August 2002
Fixed "onbelt" in T::C::Thread so that it won't hang if the belt is
shut down. Needed for Thread::Pool and in general useful.
5 August 2002
Added index functionality to "peek" and "peek_dontwait" to all
sub-classes and added documentation for it. Needed for Thread::Pool.
Finally got the T::C::Thread implementation getting through the
test-suite reliably. Problem was caused by my misconception that
cond_wait would always regain the lock() after being signalled. It
does not. This should have been in big red letters in the
threads::shared pod.
Added methods "shutdown", "thread" and "tid" plus documentation.
3 August 2002
Adapted test-suite to use many more combinations of possible
settings.
2 August 2002
Added new "optimize" field parameter to allow you to select an
implementation optimized for speed (T::C::Array and T::C::Throttled)
or for memory (T::C::Thread).
Added new implementation of Thread::Conveyor based on using a
seperate thread for keeping the belt: Thread::Conveyor::Thread.
Move basic parts of Thread::Conveyor to Thread::Conveyor::Array.
Which should allow for other implementations of the conveyor
metaphor.
0.04 1 August 2002
Added internal method "_belt" to get at the actual belt, for both
the normal and the throttled belt.
0.03 31 July 2002
Added internal methods "_freeze" and "_thaw" so that we can move
to a different serialization scheme or other optimizations without
having to change all the sub-classes as well. First user is
Thread::Conveyor::Monitored.
Added "maxjobs" and "minjobs" method in Thread::Conveyor that will
die when they're called (throttling only takes place in
Thread::Conveyor::Throttled).
Fixed problem that would cause throttling to not be switched off
when the "maxjobs" method was specified with an undef value.
Added methods "clean" and "clean_dontwait" for cleaning the belt
to both Thread::Conveyor and Thread::Conveyor::Throttled. Also
added documentation for it.
Added internal method "_clean" for obtaining frozen contents of
the belt to both Thread::Conveyor and Thread::Conveyor::Throttled.
This was needed for Thread::Conveyor::Monitored.
0.02 30 July 2002
Adapted Thread::Conveyor to default to a throttled conveyor belt
automatically if no specific parameters specified. Now allows for
specification of throttling settings in a parameter hash reference.
If unthrottling is specifically specified, returns to the old
unthrottled behaviour (where the object is just an array reference),
thus keeping the old execution speed.
Added helper class Thread::Conveyor::Throttled to allow for throttled
conveyor belts that block if there are too many boxes on the belt.
Fixed some documentation nits.
0.01 25 July 2002
First version of Thread::Conveyor, which started life as
Thread::Queue::Any.
|
