ultimatepp/uppsrc/Core/src.tpp/CoWork_en-us.tpp
cxl 5c351db10d ide: LayDes improvements, fixed TextDiff highlighting issue (light them, dark text highlighting -> invisible text)
git-svn-id: svn://ultimatepp.org/upp/trunk@14028 f0d560ea-af0d-0410-9eb7-867de7ffcac7
2020-02-18 09:06:47 +00:00

267 lines
No EOL
13 KiB
C++

topic "CoWork";
[i448;a25;kKO9;2 $$1,0#37138531426314131252341829483380:class]
[l288;2 $$2,0#27521748481378242620020725143825:desc]
[0 $$3,0#96390100711032703541132217272105:end]
[H6;0 $$4,0#05600065144404261032431302351956:begin]
[i448;a25;kKO9;2 $$5,0#37138531426314131252341829483370:item]
[l288;a4;*@5;1 $$6,6#70004532496200323422659154056402:requirement]
[l288;i1121;b17;O9;~~~.1408;2 $$7,0#10431211400427159095818037425705:param]
[i448;b42;O9;2 $$8,8#61672508125594000341940100500538:tparam]
[b42;2 $$9,9#13035079074754324216151401829390:normal]
[2 $$0,0#00000000000000000000000000000000:Default]
[{_}
[s0; &]
[ {{10000@(113.42.0) [s0;%% [*@7;4 CoWork]]}}&]
[s3; &]
[s1;:CoWork`:`:class: [@(0.0.255)3 class][3 _][*3 CoWork]&]
[s9;%% This class is indented as general parallelization tool. Whenever
jobs (e.g. loop iterations) are independent (they do not share
any data between iterations), CoWork can be used to relatively
easily spawn loop iterations over threads and thus over CPU cores.
Note that previous statement does [* not] preclude CoWork iterations
to share data at all `- sharing data using Mutex or similar serialization
mechanisms still works. CoWork works with fixed`-size thread
pool, which is created during initialization phase (which first
CoWork constructor is called). No more thread are created or
destroyed during normal work. Nesting of CoWork instances is
also possible. Thread pool is normally terminated when the main
thread finishes.&]
[s9;%% No synchronization is required to access CoWork instances
from various threads (CoWork is internally synchronized).&]
[s9;%% If an exception is thrown in worker thread, which is not handled
by worker thread, it is caught and rethrown in CoWork thread
in Finish routine. Any such exception also causes the Cancel
of the CoWork. For this reason, CoWork destructor [*/ CAN] throw
exceptions and CoWork should be usually used as automatick (stack)
variable. If you need CoWork that does not throw exceptions in
destructor, use CoWorkNX.&]
[s9;%% [*/ Implementation notes: ]Current implementation has single
global FIFO stack for 2048 scheduled jobs. When there is no slot
available when scheduling the job, it is performed immediately
by Do. Finish method has to wait until all jobs scheduled by
CoWork instance are finished, while waiting it attempts to perform
scheduled jobs from the same instance. That way work always progresses
even if there is shortage of worker threads.&]
[s0;%% &]
[ {{10000F(128)G(128)@1 [s0;%% [* Public Method List]]}}&]
[s3; &]
[s5;:Upp`:`:CoWork`:`:TrySchedule`(Upp`:`:Function`<void`(`)`>`&`&`): [@(0.0.255) stati
c bool]_[* TrySchedule]([_^Upp`:`:Function^ Function]<[@(0.0.255) void]_()>`&`&_[*@3 fn])
&]
[s5;:Upp`:`:CoWork`:`:TrySchedule`(const Upp`:`:Function`<void`(`)`>`&`): [@(0.0.255) s
tatic] [@(0.0.255) bool]_[* TrySchedule]([@(0.0.255) const]_[_^Upp`:`:Function^ Function]<
[@(0.0.255) void]_()>`&_[*@3 fn])&]
[s2;%% This is a low`-level function that attempts to schedule [%-*@3 fn]
to be executed by worker thread. Returns true if [%-*@3 fn] was
scheduled, false if not (in case there is no slot left in scheduling
stacks). Note that this function only schedules the function,
the exact time of execution is unknown.&]
[s3;%% &]
[s4; &]
[s5;:Upp`:`:CoWork`:`:Schedule`(Upp`:`:Function`<void`(`)`>`&`&`): [@(0.0.255) static
void]_[* Schedule]([_^Upp`:`:Function^ Function]<[@(0.0.255) void]_()>`&`&_[*@3 fn])&]
[s5;:Upp`:`:CoWork`:`:Schedule`(const Upp`:`:Function`<void`(`)`>`&`): [@(0.0.255) stat
ic] [@(0.0.255) void]_[* Schedule]([@(0.0.255) const]_[_^Upp`:`:Function^ Function]<[@(0.0.255) v
oid]_()>`&_[*@3 fn])&]
[s2;%% Similar to TrySchedule, but always schedules [%-*@3 fn] `- even
if it has to wait for slot to become available.&]
[s3;%% &]
[s4; &]
[s5;:Upp`:`:CoWork`:`:Do`(Upp`:`:Function`<void`(`)`>`&`&`): [@(0.0.255) void]_[* Do]([_^Upp`:`:Function^ F
unction]<[@(0.0.255) void]_()>`&`&_[*@3 fn])&]
[s5;:Upp`:`:CoWork`:`:Do`(const Upp`:`:Function`<void`(`)`>`&`): [@(0.0.255) void]_[* Do](
[@(0.0.255) const]_[_^Upp`:`:Function^ Function]<[@(0.0.255) void]_()>`&_[*@3 fn])&]
[s5;:Upp`:`:CoWork`:`:operator`&`(const Upp`:`:Function`<void`(`)`>`&`): [_^Upp`:`:CoWork^ C
oWork][@(0.0.255) `&]_[* operator`&]([@(0.0.255) const]_[_^Upp`:`:Function^ Function]<[@(0.0.255) v
oid]_()>`&_[*@3 fn])&]
[s5;:Upp`:`:CoWork`:`:operator`&`(Upp`:`:Function`<void`(`)`>`&`&`): [_^Upp`:`:CoWork^ C
oWork][@(0.0.255) `&]_[* operator`&]([_^Upp`:`:Function^ Function]<[@(0.0.255) void]_()>`&
`&_[*@3 fn])&]
[s2;%% Schedules [%-*@3 fn] to be executed. All changes to data done
before Do are visible in the scheduled code. The order of execution
or whether the code is execute in another or calling thread is
not specified. In certain situations (no scheduling slot available),
Do can perform scheduled job immediately in calling thread.&]
[s3;%% &]
[s4; &]
[s5;:Upp`:`:CoWork`:`:Loop`(Upp`:`:Function`<void`(`)`>`&`&`): [@(0.0.255) void]_[* Loop](
[_^Upp`:`:Function^ Function]<[@(0.0.255) void]_()>`&`&_[*@3 fn])&]
[s5;:Upp`:`:CoWork`:`:Loop`(const Upp`:`:Function`<void`(`)`>`&`): [@(0.0.255) void]_[* L
oop]([@(0.0.255) const]_[_^Upp`:`:Function^ Function]<[@(0.0.255) void]_()>`&_[*@3 fn])&]
[s5;:Upp`:`:CoWork`:`:operator`*`(const Upp`:`:Function`<void`(`)`>`&`): [_^Upp`:`:CoWork^ C
oWork][@(0.0.255) `&]_[* operator`*]([@(0.0.255) const]_[_^Upp`:`:Function^ Function]<[@(0.0.255) v
oid]_()>`&_[*@3 fn])&]
[s5;:Upp`:`:CoWork`:`:operator`*`(Upp`:`:Function`<void`(`)`>`&`&`): [_^Upp`:`:CoWork^ C
oWork][@(0.0.255) `&]_[* operator`*]([_^Upp`:`:Function^ Function]<[@(0.0.255) void]_()>`&
`&_[*@3 fn])&]
[s2;%% Schedules [%-*@3 fn] to be run on all worker threads and on
calling thread. After the first thread returns from [%-*@3 fn],
all other scheduled [%-*@3 fn] jobs that has not started yet are
unscheduled. Waits for all started jobs to finish. The function
also sets internal index counter to zero in CoWork before starting
any worker thread. Worker thread should acquire job quantum in
internal loop `- internal CoWork index can be used for this purpose.&]
[s3;%% &]
[s4; &]
[s5;:Upp`:`:CoWork`:`:Next`(`): [@(0.0.255) int]_[* Next]()&]
[s2;%% Atomically increments internal index counter and returns its
previous value (thus the first value returned is 0). Supposed
to be used with Loop.&]
[s3; &]
[s4; &]
[s5;:Upp`:`:CoWork`:`:FinLock`(`): [@(0.0.255) static] [@(0.0.255) void]_[* FinLock]()&]
[s2;%% This functions is to be called in scheduled routine. Its purpose
is to serialize access to shared data at the end of the routine.
The rationale is that CoWork has to lock some mutex anyway after
scheduled code finishes, so FinLock can lock this mutex a bit
earlier, joining two mutex locks into single one. Of course,
as with all locks, execution of locked code should be short as
not to cause congestion of CoWork scheduling.&]
[s3;%% &]
[s4; &]
[s5;:Upp`:`:CoWork`:`:Cancel`(`): [@(0.0.255) void]_[* Cancel]()&]
[s2;%% Removed all jobs scheduled by this thread that has not started
yet from the queue and then waits for any jobs already started
to finish. Rethrows the exception thrown in worker threads. If
more than single worker thread throws the exception, the first
exception thrown in absolute time is rethrown.&]
[s3; &]
[s4; &]
[s5;:Upp`:`:CoWork`:`:IsCanceled`(`): [@(0.0.255) static] [@(0.0.255) bool]_[* IsCanceled](
)&]
[s2;%% This methods returns true in worker thread when the worker
thread is a part of some CoWork instance and that instance was
canceled.&]
[s3; &]
[s4; &]
[s5;:CoWork`:`:Finish`(`): [@(0.0.255) void]_[* Finish]()&]
[s2;%% Waits until all jobs scheduled using Do (or operator`&) are
finished. All changes to data performed by scheduled threads
are visible after Finish. While waiting, Finish can perform scheduled
jobs. Can &]
[s3; &]
[s4; &]
[s5;:Upp`:`:CoWork`:`:IsFinished`(`): [@(0.0.255) bool]_[* IsFinished]()&]
[s2;%% Checkes whether all jobs scheduled using Do (or operator`&)
are finished. All changes to data performed by scheduled threads
are visible after IsFinished returns true (so this is basically
non`-blocking variant of Finish).&]
[s3;%% &]
[s4; &]
[s5;:Upp`:`:CoWork`:`:Reset`(`): [@(0.0.255) void]_[* Reset]()&]
[s2;%% Calls Cancel, catches and ignores all exceptions eventually
thrown by worker threads. Then resets CoWork to the initial state
as if it was just constructed. Useful when using CoWork as nonlocal
variable.&]
[s3; &]
[s4; &]
[s5;:CoWork`:`:`~CoWork`(`): [@(0.0.255) `~][* CoWork]()&]
[s2;%% Calls Finish(). Can eventually rethrow worker thread exception.
If there is a chance of destructor being involved in stack unwinding,
Finish should be called separately before destructor.&]
[s3; &]
[s4; &]
[s5;:Upp`:`:CoWork`:`:IsWorker`(`): [@(0.0.255) static] [@(0.0.255) bool]_[* IsWorker]()&]
[s2;%% Returns true if current thread is CoWork worker thread.&]
[s3; &]
[s4; &]
[s5;:Upp`:`:CoWork`:`:GetWorkerIndex`(`): [@(0.0.255) static] [@(0.0.255) int]_[* GetWorker
Index]()&]
[s2;%% Returns the index of worker `- index is >`= 0 and < GetPoolSize().
This is useful if there is a need for per`-thread resources.
`-1 means that thread is now worker (this can happen when Finish
is using calling thread to perform jobs).&]
[s3; &]
[s4; &]
[s5;:Upp`:`:CoWork`:`:GetPoolSize`(`): [@(0.0.255) static] [@(0.0.255) int]_[* GetPoolSize](
)&]
[s2;%% Returns the current count of worker threads.&]
[s3; &]
[s4; &]
[s5;:Upp`:`:CoWork`:`:SetPoolSize`(int`): [@(0.0.255) static void]_[* SetPoolSize]([@(0.0.255) i
nt]_[*@3 n])&]
[s2;%% Adjusts the thread pool size (default pool size is CPU`_Cores()
`+ 2).&]
[s0; &]
[s0; &]
[ {{10000@(113.42.0) [s0;%% [*@7;4 CoWorkNX]]}}&]
[s0; &]
[s1;:Upp`:`:CoWorkNX`:`:struct: [@(0.0.255)3 struct][3 _][*3 CoWorkNX][3 _:_][@(0.0.255)3 public
][3 _][*@3;3 CoWork]&]
[s2;%% This simple helper class just removes `"noexcept(false)`"
from CoWork destructor, removing exception throws from the destructor
and allowing it to be used as member variable. Obviously, if
exception handling is required, you need to use Finish (not depend
on destructor to call it).&]
[s0; &]
[s0; &]
[ {{10000@(113.42.0) [s0;%% [*@7;4 Loop parallelisation functions]]}}&]
[s3; &]
[s5;:Upp`:`:CoDo`(Upp`:`:Function`<void`(`)`>`&`&`): [@(0.0.255) void]_[* CoDo]([_^Upp`:`:Function^ F
unction]<[@(0.0.255) void]_()>`&`&_[*@3 fn])&]
[s2;%% Creates CoWork and calls Loop([%-*@3 fn]). This is the most
effective way to parallelise iteration.&]
[s3;%% &]
[s4; &]
[s5;:Upp`:`:CoFor`(int`,Fn`): [@(0.0.255) template]_<[@(0.0.255) typename]_[*@4 Fn]>_[@(0.0.255) v
oid]_[* CoFor]([@(0.0.255) int]_[*@3 n], [*@4 Fn]_[*@3 iterator])&]
[s2;%% Based on CoDo, runs in parallel [%-*@3 iterator] for values
0..[%-*@3 n] passing the value as argument.&]
[s0; &]
[ {{10000@(113.42.0) [s0;%% [*@7;4 CoWorkerResources]]}}&]
[s3; &]
[s1;:noref: [@(0.0.255)3 template][3 _<][@(0.0.255)3 class][3 _][*@4;3 T][3 >]&]
[s1;:Upp`:`:CoWorkerResources`:`:class: [@(0.0.255) class]_[* CoWorkerResources]&]
[s2;%% This is a simple helper class that provides per`-worker resources.
For example, certain calculation requires the instance of computation
model that is not immutable, but can be reused over iterations.
In single`-threaded code, a single instance of such model would
be used over the whole loop, however in multi`-threaded code,
each worker thread, plus thread that created CoWork need its
own instance. CoWorkerResources helps to manage such situation.&]
[s0;%% &]
[s2;%% NOTE: The problem that this helper class solves is in majority
better solved by using CoDo function.&]
[s0;%% &]
[s3;%% &]
[ {{10000F(128)G(128)@1 [s0;%% [* Public Method List]]}}&]
[s3; &]
[s5;:Upp`:`:CoWorkerResources`:`:CoWorkerResources`(`): [* CoWorkerResources]()&]
[s2;%% Creates a required number of instances so that each sub`-job
of CoWork has its unique instance.&]
[s3; &]
[s4; &]
[s5;:Upp`:`:CoWorkerResources`:`:CoWorkerResources`(Upp`:`:Event`<T`&`>`): [* CoWorkerR
esources]([_^Upp`:`:Event^ Event]<[*@4 T][@(0.0.255) `&]>_[*@3 initializer])&]
[s2;%% Creates a required number of instances so that each sub`-job
of CoWork has its unique instance.and initializes them using
[%-*@3 initializer].&]
[s3; &]
[s4; &]
[s5;:Upp`:`:CoWorkerResources`:`:GetCount`(`)const: [@(0.0.255) int]_[* GetCount]()_[@(0.0.255) c
onst]&]
[s2;%% Returns the number of instances. Note that this is equal to
CoWork`::GetPoolSize() if thread that created CoWorkerResources
is itself CoWork worker, or CoWork`::GetPoolSize() `+ 1 if it
is any other thread. The reason for this is that CoWork owner
can execute CoWork jobs too (while waiting in Finish).&]
[s3;%% &]
[s4; &]
[s5;:Upp`:`:CoWorkerResources`:`:operator`[`]`(int`): [*@4 T][@(0.0.255) `&]_[* operator`[`]
]([@(0.0.255) int]_[*@3 i])&]
[s2;%% Returns instance [%-*@3 i]. Together with GetCount can be used
to initialize resources (alternative method to using constructor
with [%-*@3 initializer].&]
[s3;%% &]
[s4; &]
[s5;:Upp`:`:CoWorkerResources`:`:Get`(`): [*@4 T][@(0.0.255) `&]_[* Get]()&]
[s5;:Upp`:`:CoWorkerResources`:`:operator`~`(`): [*@4 T][@(0.0.255) `&]_[* operator`~]()&]
[s2;%% Supposed to be called in CoWork job, returns a reference to
resources unique for the thread.&]
[s3; &]
[s4; &]
[s5;:Upp`:`:CoWorkerResources`:`:begin`(`): [*@4 T]_`*[* begin]()&]
[s5;:Upp`:`:CoWorkerResources`:`:end`(`): [*@4 T]_`*[* end]()&]
[s2;%% Standard iterator access.&]
[s3; &]
[s0; ]]