After an extended break due to "real life issues"....
Pick up the investigation, with the goal to ascertain a valid definition
and understanding of all test parameters. A first step is to establish
a baseline ''without using a computational load''; this might be some kind
of base overhead of the scheduler.
However -- the way the test scaffolding was built, it is difficult to
create a feedback loop for the statistical test setup with binary search,
since it is not really clear how the single control parameter of the test algorithm,
the so called "stress factor", shall be interpreted and how it can be
combined with a base load.
An extended series of tests, while watching the observed value patterns qualitatively,
seems to corroborate the former results, indicating that the base expense
in my test setup (using a debug build) is at ~200µs / Node / core.
Yet the difficulty to interpret this result and arrive at a logical and generic model
prevents me from translating this into a measurement scheme, which can
be executed independently from a specific test setup and hardware
While the idea with capturing observation values is nice,
it definitively does not belong into a library impl of the
search algorithm, because this is usage specific and grossly
complicates the invocation.
Rather, observation data can be captured by side-effect
from the probe-λ holding the actual measurement run.
...based on the adapted time-factor sequence
implemented yesterday in TestChainLoad itself
- in this case, the TimeBase from the computation load is used as level speed
- this »base beat« is then modulated by the timing factor sequence
- working in an additional stress factor to press the schedule uniformly
- actual start time will be added as offset once the actual test commences
...so IterExplorer got yet another processing layer,
which uses the grouping mechanics developed yesterday,
but is freely configurable through λ-Functions.
At actual usage sit in TestChainLoad, now only the actual
aggregation computation must be supplied, and follow-up computations
can now be chained up easily as further transformation layers.
...causing the system to freeze due to excess memory allocation.
Fortunately it turned out this was not an error in the Scheduler core
or memory manager, but rather a sloppiness in the test scaffolding.
However, this incident highlights that the memory manager lacks some
sanity checks to prevent outright nonsensical allocation requests.
Moreover it became clear again that the allocation happens ''already before''
entering the Scheduler — and thus the existing sanity check comes too late.
Now I've used the same reasoning also for additional checks in the allocator,
limiting the Epoch increment to 3000 and the total memory allocation to 8GiB
Talking of Gibitbytes...
indeed we could use a shorthand notation for that purpose...
The last round of refactorings yielded significant improvements
- parallelisation now works as expected
- processing progresses closer to the schedule
- run time was reduced
The processing load for this test is tuned in a way to overload the
scheduler massively at the end -- the result must be correct non the less.
There was one notable glitch with an assertion failure from the memory manager.
Hopefully I can reproduce this by pressing and overloading the Scheduler more...
..initial gauging is a tricky subject,
since existing computer's performance spans a wide scale
Allowing
- pre calibration -98% .. +190%
- single run ±20%
- benchmark ±5%
...which can be deliberately attached (or not attached) to the
individual node invocation functor, allowing to study the effect
of actual load vs. zero-load and worker contention
...during development of the Chain-Load, it became clear that we'll often
need a collection of small trees rather than one huge graph. Thus a rule
for pruning nodes and finishing graphs was added. This has the consequence
that there might now be several exit nodes scattered all over the graph;
we still want one single global hash value to verify computations,
thus those exit hashes must now be picked up from the nodes and
combined into a single value.
All existing hash values hard coded into tests must be updated
Invent a special JobFunctor...
- can be created / bound from a λ
- self-manages its storage on the heap
- can be invoked once, then discards itself
Intention is to pass such one-time actions to the Scheduler
to cause some ad-hoc transitions tied to curren circumstances;
a notable example will be the callback after load-test completion.
... so this (finally) is the missing cornerstone
... traverse the calculation graph and generate render jobs
... provide a chunk-wise pre-planning of the next batch
... use a future to block the (test) thread until completed
- decided to abstract the scheduler invocations as λ
- so this functor contains the bare loop logic
Investigation regarding hash-framework:
It turns out that boost::hash uses a different hash_combine,
than what we have extracted/duplicated in lib/hash-value.hpp
(either this was a mistake, or boost::hash did use this weaker
function at that time and supplied a dedicated 64bit implementation later)
Anyway, should use boost::hash for the time being
maybe also fix the duplicated impl in lib/hash-value.hpp
- use a dedicated context "dropped off" the TestChainLoad instance
- encode the node-idx into the InvocationInstanceID
- build an invocation- and a planning-job-functor
- let planning progress over an lib::UninitialisedStorage array
- plant the ActivityTerm instances into that array as Scheduling progresses
Introduced as remedy for a long standing sloppiness:
Using a `char[]` together with `reinterpret_cast` in storage management helpers
bears danger of placing objects with wrong alignment; moreover, there are increasing
risks that modern code optimisers miss the ''backdoor access'' and might apply too
aggressive rewritings.
With C++17, there is a standard conformant way to express such a usage scheme.
* `lib::UninitialisedStorage` can now be used in a situation (e.g. as in `ExtentFamily`)
where a complete block of storage is allocated once and then subsequently used
to plant objects one by one
* moreover, I went over the code base and adapted the most relevant usages of
''placement-new into buffer'' to also include the `std::launder()` marker
... special rule to generate a fixed expansion on each seed
... consecutive reductions join everything back into one chain
... can counterbalance expansions and reductions
...as it turns out, the solution embraced first was the cleanest way
to handle dynamic configuration of parameters; just it did not work
at that time, due to the reference binding problem in the Lambdas.
Meanwhile, the latter has been resolved by relying on the LazyInit
mechanism. Thus it is now possible to abandon the manipulation by
side effect and rather require the dynamic rule to return a
''pristine instance''.
With these adjustments, it is now possible to install a rule
which expands only for some kinds of nodes; this is used here
to crate a starting point for a **reduction rule** to kick in.
It seams indicated to verify the generated connectivity
and the hash calculation and recalculation explicitly
at least for one example topology; choosing a topology
comprised of several sub-graphs, to also verify the
propagation of seed values to further start-nodes.
In order to avoid addressing nodes directly by index number,
those sub-graphs can be processed by ''grouping of nodes'';
all parts are congruent because topology is determined by
the node hashes and thus a regular pattern can be exploited.
To allow for easy processing of groups, I have developed a
simplistic grouping device within the IterExplorer framework.
- with the new pruning option, start-Nodes can now be anywhere
- introduce predicates to detect start-Nodes and exit-Nodes
- ensure each new seed node gets the global seed on graph construction
- provide functionality to re-propagate a seed and clear hashes
- provide functionality to recalculate the hashes over the graph
up to now, random values were completely determined by the
Node's hash, leading to completely symmetrical topology.
This is fine, but sometimes additional randomness is desirable,
while still keeping everything deterministic; the obvious solution
is to make the results optionally dependent on the invocation order,
which is simply to achieve with an additional state field. After some
tinkering, I decided to use the most simplistic solution, which is
just a multiplication with the state.
...so this was yet another digression, caused by the desire
somehow to salvage this problematic component design. Using a
DSL token fluently, while internally maintaining a complex and
totally open function based configuration is a bit of a stretch.
For context: I've engaged into writing a `LazyInit` helper component,
to resolve the inner contradiction between DSL use of `RandomDraw`
(implying value semantics) and the design of a processing pipeline,
which quite naturally leads to binding by reference into the enclosing
implementation.
In most cases, this change (to lazy on-demand initialisation) should be
transparent for the complete implementation code in `RandomDraw` -- with
one notable exception: when configuring an elaborate pipeline, especially
with dynamic changes of the probability profile during the simulation run,
then then obviously there is the desire to use the existing processing
pipeline from the reconfiguration function (in fact it would be quite
hard to explain why and where this should be avoided). `LazyInit` breaks
this usage scenario, since -- at the time the reconfiguration runs --
now the object is not initialised at all, but holds a »Trojan« functor,
which will trigger initialisation eventually.
After some headaches and grievances (why am I engaging into such an
elaborate solution for such an accidental and marginal topic...),
unfortunately it occurred to me that even this problem can be fixed,
with yet some further "minimal" adjustments to the scheme: the LazyInit
mechanism ''just needs to ensure'' that the init-functor ''sees the
same environment as in eager init'' -- that is, it must clear out the
»Trojan« first, and it ''could apply any previous pending init function''
fist. That is, with just a minimal change, we possibly build a chain
of init functors now, and apply them in given order, so each one
sees the state the previous one created -- as if this was just
direct eager object manipulation...
...this is a more realistic demo example, which mimics
some of the patterns present in RandomDraw. The test also
uses lambdas linking to the actual storage location, so that
the invocation would crash on a copy; LazyInit was invented
to safeguard against this, while still allowing leeway
during the initialisation phase in a DSL.
...oh my.
This is getting messy. I am way into danger territory now....
I've made a nifty cool design with automatically adapted functors;
yet at the end of the day, this does not bode well with a DSL usage,
where objects appear to be simple values from a users point of view.
- Helper function to find out of two objects are located
"close to each other" -- which can be used as heuristics
to distinguish heap vs. stack storage
- further investigation shows that libstdc++ applies the
small-object optimisation for functor up to »two slots«
in size -- but only if the copy-ctor is trivial. Thus
a lambda capturing a shared_ptr by value will *always*
be maintained in heap storage (and LazyInit must be
redesigned accordingly)...
- the verify_inlineStorage() unit test will now trigger
if some implementation does not apply small-object optimisation
under these minimal assumptions
the RandomDraw rules developed last days are meant to be used
with user-provided λ-adapters; employing these in a context
of a DSL runs danger of producing dangling references.
Attempting to resolve this fundamental problem through
late-initialisation, and then locking the component into
a fixed memory location prior to actual usage. Driven by
the goal of a self-contained component, some advanced
trickery is required -- which again indicates better
to write a library component with adequate test coverage.
...now using the reworked partial-application helper...
...bind to *this and then recursively re-invoke the adaptation process
...need also to copy-capture the previously existing mapping-function
first test seems to work now
Investigation in test setup reveals that the intended solution
for dynamic configuration of the RandomDraw can not possibly work.
The reason is: the processing function binds back into the object instance.
This implies that RandomDraw must be *non-copyable*.
So we have to go full circle.
We need a way to pass the current instance to the configuration function.
And the most obvious and clear way would be to pass it as function argument.
Which however requires to *partially apply* this function.
So -- again -- we have to resort to one of the functor utilities
written several years ago; and while doing so, we must modernise
these tools further, to support perfect forwarding and binding
of reference arguments.
- strive at complete branch coverage for the mapping function
- decide that the neutral value can deliberately lie outside
the value range, in which case the probability setting
controls the number of _value_ result incidents vs
neutral value result incidents.
- introduce a third path to define this case clearly
- implement the range setting Builder-API functions
- absorb boundrary and illegal cases
For sake of simplicity, since this whole exercise is a byproduct,
the mapping calculations are done in doubles. To get even distribution
of values and a good randomisation, it is thus necessary to break
down the size_t hash value in a first step (size_t can be 64bit
and random numbers would be subject to rounding errors otherwise)
The choice of this quantiser is tricky; it must be a power of two
to guarantee even distribution, and if chosen to close to the grid
of the result values, with lower probabilities we'd fail to cover
some of the possible result values. If chosen to large, then
of course we'd run danger of producing correlated numbers on
consecutive picks.
Attempting to use 4 bits of headroom above the log-2 of the
required value range. For example, 10-step values would use
a quantiser of 128, which looks like a good compromise.
The following tests will show how good this choice holds up.
This highly optimised function was introduced about one year ago
for handling of denomals with rational values (fractions), as
an interim solution until we'll switch to C++20.
Since this function uses an unrolled loop and basically
just does a logarithmic search for the highest set bit,
it can just be declared constexpr. Moreover, it is now
relocated into one of the basic utility headers
Remark: the primary "competitor" is the ilogb(double),
which can exploit hardware acceleration. For 64bit integers,
the ilog2() is only marginally faster according to my own
repeated invocation benchmarks.
The first step was to allow setting a minimum value,
which in theory could also be negative (at no point is the
code actually limited to unsigned values; this is rather
the default in practice).
But reconsidering this extensions, then you'd also want
the "neutral value" to be handled properly. Within context,
this means that the *probability* controls when values other
than the neutral value are produced; especially with p = 1.0
the neutral value shall not be produced at all
...since the Policy class now defines the function signature,
we can no longer assume that "input" is size_t. Rather, all
invocations must rely on the generic adaptaion scheme.
Getting this correct turns out rather tricky again;
best to rely on a generic function-composition.
Indeed I programmed such a helper several years ago,
with the caveat that at that time we used C++03 and
could not perfect-forward arguments. Today this problem
can be solved much more succinct using generic Lambdas.
to define this as a generic library component,
any reference to the actual data source moust be extracted
from the body of the implementation and supplied later
at usage site. In the actual case at hand the source
for randomness would be the node hash, and that is
absolutely an internal implementation detail.
The idea is to use some source of randomness to pick a
limited parameter value with controllable probability.
While the core of the implementation is nothing more
than some simple numeric adjustments, these turn out
to be rather intricate and obscure; the desire to
package these technicalities into a component
however necessitates to make invocations
at usage site self explanatory.
...these were developed driven by the immediate need
to visualise ''random generated computation patterns''
for ''Scheduler load testing.''
The abstraction level of this DSL is low
and structures closely match some clauses of the DOT language;
this approach may not yet be adequate to generate more complex
graph structures and was extracted as a starting point
for further refinements....
With all the preceding DSL work, this turns out to be surprisingly easy;
the only minor twist is the grouping of nodes into (time)levels,
which can be achieved with a "lagging" update from the loop body
Note: next step will be to extract the DSL helpers into a Library header
...refine the handling of FrameRates close to the definition bounds
...implement the actual rule to scale allocator capacity on announcement
...hook up into the seedCalcStream() with a default of +25FPS
+ test coverage
...whenever a new CalcStream is seeded, it would be prudent
not only to step up the WorkForce (which is already implemented),
but also to provide a hint to the BlockFlow allocator regarding
the expected calculation density.
Such a hint would allow to set a more ample »epoch« spacing,
thereby avoiding to drive the allocator into overload first.
The allocator will cope anyway and re-balance in a matter of
about 2 seconds, but avoiding this kind of control oscillations
altogether will lead to better performance at calculation start.
Use a simple destructor-trick to set up a concise notation
for temporarily manipulating a value for testing.
The manipulation will automatically be undone
when leaving scope
While testing, I repeatedly had SEGFAULT in the new thread-wrapper,
but only when running under debugger. While the language spec guarantees
that exit from the thread handle initialisation synchronizes-with
the start of the new thread, there is no guarantee in the reverse
direction. Here this means that the new thread may not see the
newly initialised thread handle ID at start. Thus I've added
a yield-wait at the very beginning of the new thread function.
Under normal conditions, the startup of a thread takes at least
100 - 500µs and thus I've never seen the problematic behaviour
without debugger. However, adding a yield-wait loop at that point
seems harmless (it typically checks back every 400ns or so).
All real usages of the thread wrapper in the application use
some kind of additional coordination or even a sync barrier
to ensure the thread can pick up all further data before
going into active work.
WARNING: if someone would detach() the thread immediately after
creating it, then this added condition would cause the starting
thread function to hang forever. In our current setup for the
thread wrapper, this is not possible, since the thread handle
is embedded into protected code. The earliest point you could
do that would be in the handle_begin_thread(), which is called
from the thread itself *after* the new check. And moreover,
this would require to write a new variation of the Policy.
While building increasingly complex integration tests for the Scheduler,
it turns out helpful to be able to manipulate the "full concurreency"
as used by Scheduler, WorkForce and LoadController.
In the current test, I am facing a problem that new entries from the
threadsafe entrance queue are not propagated to the priority queue
soon enough; partly this is due to functionality still to be added
(scaling up when new tasks are passed in) -- but this will further
complicate the test setup.
as an aside, the header lib/test/microbenchmark.hpp
turns out to be prolific for this kind of investigation.
However, it is somewhat obnoxious that the »test subject«
must expose the signature <size_t(size_t)>.
Thus, with some metaprogramming magic, an generic adaptor
can be built to accept a range of typical alternatives,
and even the quite obvious signature void(void).
Since all these will be wrapped directly into a lambda,
the optimiser will remove these adaptations altogether.
...there seemed to be an anomaly of 50...100µs
==> conclusion: this is due to the instrumentation code
- it largely caused by the EventLog, which was never meant
to be used in performance-critical code, and does hefty
heap allocations and string processing.
- moreover, there clearly is a cache-effect, adding a Factor 2
whenever some time passed since the last EventLog call
==> can be considered just an artifact of the test setup and
will have no impact on the scheduler
remark: this commit adds a lot of instrumentation code
To cover the visible behaviour of the work-Function,
we have to check an amalgam of timing delays and time differences.
This kind of test tends to be problematic, since timings are always
random and also machine dependent, and thus we need to produce pronounced effects
...to bring it more in line with all the other calls dealing with Activity*
...allows also to harmonise the ActivityLang::dispatchChain()
...and to compose the calls in Scheduler directly
NOTE: there is a twist: our string-formatting helper did not render
custom string conversions for objects passed as pointer. This was a
long standing problem, caused by ambiguous templates overloads;
now I've attempted to solve it one level more down, in util::StringConv.
This solution may turn out brittle, since we need to exclude any direct
string conversion, most notably the ones for C-Strings (const char*)
In case this solution turns out unsustainable, please feel free
to revert this API change, and return to passing Activity& in λ-post,
because in the end this is cosmetics.
The approach to provide the ExecutionCtx seems to work out well;
after some investigation I found a solution how to code a generic
signature-check for "any kind of function-like member"...
(the trick is to pass a pointer or member-pointer, which happens
to be syntactically the same and can be handled with our existing
function signature helper after some minor tweaks)
As follow-up to the rework of thread-handling, likewise also
the implementation base for locking was switched over from direct
usage of POSIX primitives to the portable wrappers available in
the C++ standard library. All usages have been reviewed and
modernised to prefer λ-functions where possible.
With this series of changes, the old threadpool implementation
and a lot of further low-level support facilities are not used
any more and can be dismantled. Due to the integration efforts
spurred by the »Playback Vertical Slice«, several questions of
architecture could be decided over the last months. The design
of the Scheduler and Engine turned out different than previously
anticipated; notably the Scheduler now covers a wider array of
functionality, including some asynchronous messaging. This has
ramifications for the organisation of work tasks and threads,
and leads to a more deterministic memory management. Resource
management will be done on a higher level, partially superseding
some of the concepts from the early phase of the Lumiera project.
This is Step-2 : change the API towards application
Notably all invocation variants to support member functions
or a reference to bool flags are retracted, since today a
λ-binding directly at usage site tends to be more readable.
The function names are harmonised with the C++ standard and
emergency shutdown in the Subsystem-Runner is rationalised.
The old thread-wrapper test is repurposed to demonstrate
the effectiveness of monitor based locking.
After the fundamental switch from POSIX to the C++14 wrappers
the existing implementation of the Monitor can now be drastically condensed,
removing several layers of indirection. Moreover, all signatures
shall be changed to blend in with the names and patterns established
by the C++ standard.
This is Step-1 : consolidate the Implementation.
(to ensure correctness, the existing API towards application code was retained)
While not directly related to the thread handling framework,
it seems indicated to clean-up this part of the application alongside.
For »everyday« locking concerns, an Object Monitor abstraction was built
several years ago and together with the thread-wrapper, both at that time
based on direct usage of POSIX. This changeset does a mere literal
replacement of the POSIX calls with the corresponding C++ wrappers
on the lowest level. The resulting code is needlessly indirect, yet
at API-level this change is totally a drop-in replacment.
The WorkForce (passive worker pool) has been coded just recently,
and -- in anticipation of this refactoring -- directly against std::thread
instead of using the old framework.
...the switch is straight-forward, using the default case
...add the ability to decorate the thread-IDs with a running counter
This solution is basically equivalent to the version implemented directly,
but uses the lifecycle-Hooks available through `ThreadHookable`
to structure the code and separate the concerns better.
This largely completes the switch to the new thread-wrapper..
**the old implementation is not referenced anymore**
Add a complete demonstration for a setup akin to what we use
for the Session thread: a threaded component which manages itself
but also exposes an external interface, which is opened/closed alongside
...extract and improve the tuple-rewriting function
...improve instance tracking test dummy objects
...complete test coverage and verify proper memory handling
After quite some detours, with this take I'm finally able to
provide a stringent design to embody all the variants of thread start
encountered in practice in the Lumiera code base.
Especially the *self-managed* thread is now represented as a special-case
of a lifecycle-hook, and can be embodied into a builder front-end,
able to work with any client-provided thread-wrapper subclass.
extract into helper function to improve legibility.
This code is rather tricky since on invocation the hook is only provided
but not invoked. Rather, to adapt the argument types, it is wrapped
into a λ for adaptation, which then must be again bound *by value*
into yet another λ, since the Launch configuration builder is comprised
of a chain of captured functors, to be invoked later from the body of the
thread-wrapper object; this indirect procedure is necessary to ensure
all members are initialised *before* the new thread starts
to cover the identified use-cases a wide variety of functors
must be accepted and adapted appropriately. A special twist arises
from the fact that the complete thread-wrapper component stack works
without RTTI; a derived class can not access the thread-wrapper internals
while the policy component to handle those hooks can not directly downcast
to some derived user provided class. But obviously at usage site it
can be expected to access both realms from such a callback.
The solution is to detect the argument type of the given functor
and to build a two step path for a safe static cast.
after some further mulling over the design, it became clear that
a rather loose coupling to the actual usage scenario is preferrable.
Thus, instead of devising a fixed scheme how to reflect the thread state,
rather the usage can directly hook into some points in the thread lifecycle.
So this policy can be reduced to provide additional storage for functon objects.
...after resolving the fundamental design problems,
a policy mix-in can be defined now for a thread that deletes
its own wrapper at the end of the thread-function.
Such a setup would allow for »fire-and-forget« threads, but with
wrapper and ensuring safe allocations. The prominent use case
for such a setup would be the GUI-Thread.
So this finally solves the fundamental problem regarding a race on
initialisation of the thread-wrapper; it does *not* solve the same problem
for classes deriving from thread-wrapper, which renders this design questionable
altogether -- but this is another story.
In the end, this initialisation-race is rooted in the very nature of starting a thread;
it seems there are the two design alternatives:
- expose the thread-creation directly to user code (offloading the responsibility)
- offer building blocks which are inherently dangerous
this is a mere rearrangement of code (+lots of comments),
but helps to structure the overall construction better.
ThreadWrapper::launchThread() now does the actual work to build
the active std::thread object and assign it to the thread handle,
while buildLauncher is defined in the context of the constructors
and deals with wiring the functors and decaying/copying of arguments.
If we package all arguments together into a single tuple,
even including the member-function reference and the this-ptr
for the invokeThreadFunction(), which is the actual thread-functor,
then we can rely on std::make_from_tuple<T>(tuple), which implements
precisely the same hand-over via a std::index_sequence, as used by the
explicitly coded solution -- getting rid of some highly technical boilerplate
Concept study of the intended solution successful.
Can now transparently embed any conceivable functor
and an arbitrary argument sequence into a launcher-λ
Materialising into a std::tuple<decay_t<TYPES...>> did the trick.
Considering a solution to shift the actual launch of the new thread
from the initialiser list into the ctor body, to circumvent the possible
"undefined behaviour". This would also be prerequisite for defining
a self-managed variant of the thread-wrapper.
Alternative / Plan.B would be to abandon the idea of a self-contained
"thread" building block, instead relying on precise setup in the usage
context -- however, not willing to yield yet, since that would be exactly
what I wanted to avoid: having technicalities of thread start, argument
handover and failure detection intermingled with the business code.
On a close look, the wrapper design as pursued here
turns out to be prone to insidious data race problems.
This was true also for the existing solution, but becomes
more clear due to the precise definitions from the C++ standard.
This is a confusing situation, because these races typically do not
materialise in practice; due to the latency of the OS scheduler the
new thread starts invoking user code at least 100µs after the Wrapper
object is fully constructed (typically more like 500µs, which is a lot)
The standard case (lib::Thread) in its current form is correct, but borderline
to undefined behaviour, and any initialisation of members in a derived class
would be off limits (the thread-wrapper should not be used as baseclass,
rather as member)
...while reworking the application code, it became clear that
actually there are two further quite distinct variants of usage.
And while these could be implemented with some trickery based on
the Thread-wrapper defined thus far, it seems prudent better to
establish a safely confined explicit setup for these cases:
- a fire-and-forget-thread, which manages its own memory autonomously
- a thread with explicit lifecycle, with detectable not-running state
FamilyMember::allocateNextMember() was actually a post-increment,
so (different than with TypedCounter) here no correction is necessary
As an asside, WorkForce_test is sometimes unstable immediately after a build.
Seemingly a headstart of 50µs is not enough to compensate for scheduler leeway
The existing TypedCounter_test was excessively clever and convoluted,
yet failed to test the critical elements systematically. Indeed, two
bugs were hidden in synchronisation and instance access.
- build a new concurrent test from scratch, now using the threadBenchmark
function for the actual concurrent execution and just invoked a
random selected access to the counter repeatedly from a large number
of threads.
- rework the TypedContext and counter to use Atomics where applicable;
measurements indicate however that this has only negligible impact
on the amortised invocation times, which are around 60ns for single-threaded
access, yet can increase by factor 100 due to contention.
...these were already written envisionaging he new API,
so it's more or less a drop-in replacement.
- cant use vector anymore, since thread objects are move-only
- use ScopedCollection instead, which also has the benefit of
allocating the requires space up-front. Allow to deduce the
type parameter of the placed elements
... which became apparent after switching to the new Thread-wrapper implementation
... the reason is a bug in the Thread-Monitor (which will also be reworked soon)
While seemingly subtle, this is a ''deep change.''
Up to now, the project attempted to maintain two mutually disjoint
systems of error reporting: C-style error flags and C++ exceptions.
Most notably, an attempt was made to keep both error states synced.
During the recent integration efforts, this increasingly turned out
as an obstacle and source for insidious problems (like deadlocks).
As a resolve, hereby the relation of both systems is **clarified**:
* C-style error flags shall only be set and used by C code henceforth
* C++ exceptions can (optionally) be thrown by retrieving the C-style error code
* but the opposite is now ''discontinued'' : Exceptions ''do not set'' the error flag anymore
- the deadlock was caused by leaking error state through the C-style lumiera_error
- but the reason for the deadlock lies in the »convenience shortcut«
in the Object-Monitor scope guard for entering a wait state immediately.
This function undermines the unlocking-guarantee, when an exception
emanates from within the wait() function itself.
...this function was also ported to the new wrapper,
and can be verified now in a much more succinct way.
''This completes porting of the thread-wrapper''
Since the decision was taken to retain support for this special feature,
and even extend it to allow passing values, the additional functionality
should be documented in the test. Doing so also highlighted subtle problems
with argument binding.
A subtle yet important point: arguments will always be copied into the new thread.
This is a (very sensible) limitation introduced by the C++ standard.
To support seamless use, the thread-wrapper now rewrites the argument types
picked up from the invocation, to prevent passing on a reference type,
which typically ensues when invoking with a variable name. Otherwise
confusing error messages would be emitted from deep within the STD library.
As a further consequence, function signatures involving reference arguments
can no longer be bound (which is desirable; a function to be performed
within a separate thread must either rely on value arguments, or deliberately
use std::ref wrappers to pass references, assuming you know what you're doing)
- it is not directly possible to provide a variadic join(args...),
due to overload resolution ambiguities
- as a remedy, simplify the invocation of stringify() for the typical cases,
and provide some frequently used shortcuts
A common usage pattern is to derive from lib::Thread
and then implement the actual thread function as a member function
of this special-Thread-object (possibly also involving other data members)
Provide a simplified invocation for this special case,
also generating the thread-id automatically from the arguments
after all this groundwork, implementing the invocation,
capturing and hand-over of results is simple, and the
thread-wrapper classes became fairly understandable.
This relieves the Thread policy from a lot of technicalities,
while also creating a generally useful tool: the ability to invoke
/anything callable/ (thanks to std::invoke) in a fail-safe way and
transform the exception into an Either type
on second thought, the ability to transport an exception still seems
worthwhile, and can be achieved by some rearrangements in the design.
As preparation, reorganise the design of the Either-wrapper (lib::Result)
VERIFY_ERROR allows to check that an expected except is actually thrown.
The implementation was lazy however;
it just investigated the C-style error flag instead of *really* verifying
that an *lumiera::Exception* with the expected flag was caught.
This discrepancy can be a problem when there is a stray error flag set,
or for some reason the error flag gets cleared before the exception
reaches the top-level catch-block in the test.
- relocate some code into a dedicated translation unit to reduce #includes
- actually set the thread-ID (the old implementation had only a TODO at that point)
While it would be straight forward from an implementation POV
to just expose both variants on the API (as the C++ standard does),
it seems prudent to enforce the distinction, and to highlight the
auto-detaching behaviour as the preferred standard case.
Creating worker threads just for one computation and joining the results
seemed like a good idea 30 years ago; today we prefer Futures or asynchronous
messaging to achieve similar results in a robust and performant way.
ThreadJoinable can come in handy however for writing unit tests, were
the controlling master thread has to wait prior to perform verification.
So the old design seems well advised in this respect and will be retained
- cut the ties to the old POSIX-based custom threadpool framework
- remove operations deemed no longer necessary
- sync() obsoleted by the new SyncBarrier
- support anything std::invoke supports
...which is the technique used in the existing Threadpool framwork.
As expected, such a solution is significantly slower than the new
atomics-based implementation. Yet how much slower is still striking.
Timing measurements in concurrent usage situation.
Observed delay is in the order of magnitude of known scheduling leeway;
assuming thus no relevant overhead related to implementation technique
Over time, a collection of microbenchmark helper functions was
extracted from occasional use -- including a variant to perform
parallelised microbenchmarks. While not used beyond sporadic experiments yet,
this framework seems a perfect fit for measuring the SyncBarrier performance.
There is only one catch:
- it uses the old Threadpool + POSIX thread support
- these require the Threadpool service to be started...
- which in turn prohibits using them for libary tests
And last but not least: this setup already requires a barrier.
==> switch the existing microbenchmark setup to c++17 threads preliminarily
(until the thread-wrapper has been reworked).
==> also introduce the new SyncBarrier here immediately
==> use this as a validation test of the setup + SyncBarrier
Using the same building blocks, this operation can be generalised even more,
leading to a much cleaner implementation (also with better type deduction).
The feature actually used here, namely summing up all values,
can then be provided as a convenience shortcut, filling in std::plus
as a default reduction operator.
...first used as part of the test harness;
seemingly this is a generic and generally useful shortcut,
similar to algorithm::reduce (or some kind of fold-left operation)
Intended as replacement for the Mutex/ConditionVar based barrier
built into the exiting Lumiera thread handling framework and used
to ensure safe hand-over of a bound functor into the starting new
thread. The standard requires a comparable guarantee for the C++17
concurrency framework, expressed as a "synchronizes_with" assertion
along the lines of the Atomics framework.
While in most cases dedicated synchronisation is thus not required
anymore when swtiching to C++17, some special extended use cases
remain to be addressed, where the complete initialisation of
further support framework must be ensured.
With C++20 this would be easy to achieve with a std::latch, so we
need a simple workaround for the time being. After consideration of
the typical use case, I am aiming at a middle ground in terms of
performance, by using a yield-wait until satisfying the latch condition.
The investigation for #1279 leads to the following conclusions
- the features and the design of our custom thread-wrapper
almost entirely matches the design chosen meanwhile by the C++ committee
- the implementation provided by the standard library however uses
modern techniques (especially Atomics) and is more precisely worked out
than our custom implementation was.
- we do not need an *active* threadpool with work-assignment,
rather we'll use *active* workers and a *passive* pool,
which was easy to implement based on C++17 features
==> decision to drop our POSIX based custom implementation
and to retrofit the Thread-wrapper as a drop-in replacement
+++ start this refactoring by moving code into the Library
+++ create a copy of the Threadwrapper-code to build and test
the refactorings while the application itself still uses
existing code, until the transition is complete
Up to now, the DiagnosticFun mock in ActivityDetector only
created an EventLog entry on invocation and was able to retunr
a canned result value. Yet for the job invocation scenario test,
it would be desirable to hook-in a λ with a fake implementation
into the ExecutionContext. As a further convenience, the
return value is now default initialised, instead of being
marked as uninitialised until invocation of "returning(val)"
...turns out that util::toString does not explicitly handle pointers differently,
for very good reasons; this function must always work, always produce a simple and
compact representation, and it must be possible to instantiate the template
and take a function reference (which precludes adding an overload for pointers)
requires to supplement EventLog matching primitives
to pick and verify a specific positional argument.
Moreover, it is more or less arbitrary which job invocation parameters
are unpacked and exposed for verification; we'll have to see what is
actually required for writing tests...
Testcase (detect function invocation) passes now as expected
Some Library / Framework changes
- rename event-log-test.cpp
- allow the ExpectString also to work with concatenated expectation strings
Remark: there was a warning in the comment in event-log.hpp,
pointing out that negative assertions are shallow.
However, after the rework in 9/2018 (commit: d923138d1)
...this should no longer be true, since we perform proper backtracking,
leading to an exhaustive search.
ActivityMatch inherits privately from the EventMatch object,
and is thus able to delegate relevant matching queries, but
also to provide high-level special matchers.
This new design resolves the ambiguity regarding function arguments.
Moreover, we can now record the current sequence-Number as *attribute*
in the respective log record (this is the benefit of using structured
log entries instead of just a textual log), thereby avoiding the various
pitfalls with explicit bracketing sequence-number log entries
bottom line: this reworked design seems to be a better fit,
even while technically the implementation with the wrapped matcher
is somewhat ugly...
The EventLog seems to provide all the building blocks, but we need
some higher level special matchers (and maybe we also want to hide
some of the basic EventLog matchers). A soulution might be to wrap
the EventMatcher and delegate all follow-up builder calls.
This seems adequate, since the EventLog-Matcher is basically used as black box,
building up more elaborate matchers from the provided basic matchers...
Spent some time again to understand how EventLog matching works.
My feelings towards this piece of code are always the same: it is
somewhat too "tricky", but I am not aware of any other technique
to get this degree of elaborate chained matching on structured records,
short of building a dedicated matching engine from scratch.
The other alternative would be to use a flat textual log (instead of
the structured log records from EventLog), but then we'd have to
generate quite intricate regular expressions from the builder,
and I'm really doubtful it would be easier and clearer....
...turns out this is entirely generic and not tied to the context
within ActivityDetector, where it was first introduced to build a
mock functor to log all invocations.
Basically this meta-function generates a new instantiation of the
template X, using the variadic argument pack from template U<ARGS...>
...for coverage of the Activity-Language,
various invocations of unspecific functions must be verified,
with the additional twist that the implementation avoids indirections
and is thus hard to rig for tests.
Solution-Idea: provide a λ-mock to log any invocation into the
Event-Log helper, which was created some years ago to trace GUI communication...
Further extensive testing with parameter variations,
using the test setup in `BlockFlow_test::storageFlow()`
- Tweaks to improve convergence under extreme overload;
sudden load peaks are now accomodated typically < 5 sec
- Make the test definition parametric, to simplify variations
- Extract the generic microbenchmark helper function
- Documentation
- BUG: must prevent the Epoch size to become excessive low
- Problem: feedback signal should not be overly aggressive
Fine-Tuning:
- Dose for Overflow-compensation is delicate
- Moving average and Overflow should be balanced
- ideally the compensatory actions should be one order of magnitude
slower than the characteristic regulation time
Improvement: perform Moving-Average calculations in doubles
- fix a bug in IterExplorer: when iterating a »state core« directly,
the helper CoreYield passed the detected type through ValueTypeBindings.
This is logically wrong, because we never want to pick up some typedefs,
rather we always want to use the type directly returned from CORE::yield()
Here the iterator returns an Epoch&, which itself is again iterable
(it inherits from std::array<Activity, N>). However, it is clear
that we must not descent into such a "flatMap" style recursive expansion
- draft a simple scheme how to regulate Epoch lengths dynamically
- add diagnostics to pinpoint a given Activity and find out into which
Epoch it has been allocated; used to cover the allocator behaviour
Especially for the BlockFlow allocator, sanity checks are elided
for performance reasons; yet, generally speaking, it can be a very bad idea
to "optimise" away sanity checks. Thus an additional adaptor is provided
to layer such checks on top of an existing core; and IterEplorer now
always wires in this additional adaptor, and so the original behaviour
is now restored in this respect (and for the largest part of the code base)
While at first sight just a superficial variation of the existing IterStateWrapper,
it became clear with the evolution of the IterExplorer framework that
this setup represents a distinct concept, and especially lends itself
for complex and cohesive collaboration in a layered pipeline. Which
may, or may not be a good idea, depending on the circumstances.
Now, for the implementation of the scheduler memory allocation scheme,
another twist is added to the picture: we can not effort the sanity checks
on each access, even more so when layering / adapting iterators, where
it is essential that the optimiser can remove all unnecessary warts.
Library: add "obvious" utility to the IterExplorer, allowing to
materialise all contents of the Pipeline into a container
...use this to take a snapshot of all currently active Extent addresses
Using a Storage* within a wrapper as "pos" will work,
but is borderline trickery, since it amounts to subverting
the idea behind IterAdapter (which is to encapsulate a target
pointer with some control-logic in the managing container).
Using the same storage size and implementation overhead,
it is much more straight-forward to package the complete
iteration logic into a »State Core«, which in this case
however maintains a back-link to the ExtentFamily.
Iteration should just yield an Reference to an Extent,
thereby hiding all details of the actual raw storage (char[]).
This can be achieved by usind a wrapper type around a pointer
into the managing vector; from this pointer we may convert
into a vector::iterator with the trick described here
https://stackoverflow.com/a/37101607/444796
Furthermore, continued planning of the Activity-Language,
basically clarified the complete usage scenario for now;
seems all implementable right away without further difficulties
- the idea is to use slot-0 in each extent for administrative metadata
- to that end, a specialised GATE-Activity is placed into slot-0
- decision to use the next-pointer for managing the next free slot
- thus we need the help of the underlying ExtentFamily for navigating Extents
Decision to refrain from any attempt to "fix" excessive memory usage,
caused by Epochs still blocked by pending IO operations. Rather, we
assume the engine uses sane parametrisation (possibly with dynamic adjustment)
Yet still there will be some safety limit, but when exceeding this limit,
the allocator will just throw, thereby killing the playback/render process
The second design from 2017, based on a pipeline builder,
is now renamed `TreeExplorer` ⟼ `IterExplorer` and uses
the memorable entrance point `lib::explore(<seq>)`
✔
after completing the recent clean-up and refactoring work,
the monad based framework for recursive tree expansion
can be abandoned and retracted.
This approach from functional programming leads to code,
which is ''cool to write'' yet ''hard to understand.''
A second design attempt was based on the pipeline and decorator pattern
and integrates the monadic expansion as a special case, used here to
discover the prerequisites for a render job. This turned out to be
more effective and prolific and became standard for several exploring
and backtracking algorithms in Lumiera.
This very deep change (which requires almost complete rebuild)
was prompted by the need to process an object (JobPlanning),
which holds several references and is thus move-only, in the
middle of a complex processing pipeline with child expansion.
If this works out well, a long-standing and obnoxious problem
with transforming iterators would be solved, albeit by incurring
a (presumably small) performance overhead, since now the new
value is no longer *assigned*, but rather the existing payload
is destroyed and a new instance is copy/move constructed into
the inline buffer.
The primary purpose (and widely used in Lumieara) is to have a
Lambda create a new Object, which is then returned by value
and thus immediately moved into this inline buffer, where it
resides for further use (as long as the enclosing pipeline
stays alive). Unless such an object does very elaborate
allocations and registrations behind the scene, the
expense of assigning vs creating should be the same.
- had to fix a logical inconsistency in the underlying Expander implementation
in TreeExplorer: the source-pipeline was pulled in advance on expansion,
in order to "consume" the expanded element immediately; now we retain
this element (actually inaccessible) until all of the immediate
children are consumed; thus the (visible) state of the PipeFrameTick
stays at the frame number corresponding to the top-level frame Job,
while possibly expanding a complete tree of flexible prerequisites
This test now gives a nice visualisation of the interconnected states
in the Job-Planning pipeline. This can be quite complex, yet I still think
that this semi-functional approach with a stateful pipeline and expand functors
is the cleanest way to handle this while encapsulating all details
`steam/engine/mock-dispatcher.hpp |cpp` now integrates this
''complete mock setup for render jobs and frame dispatching.''
The exising `DummyJob` has been slightly adapted and renamed
to `MockJob` and is tightly integrated with the other mocks.
The implementation of a `MockDispatcher` necessitated to change
the use of `MockJobTicket`. The initial attempts used a complete
mock implementation, but this approach turned out not to be viable.
Instead — based on the ideas developed for the mock setup —
now the prospective real implementation of `JobTicket` is available
and will be used by the mock setup too. Instead of a synthetic spec,
now a setup of recursively connected `ExitNode`(s) is used; the latter
seems to develop into some kind of Facade for the render node network.
Based on this mock setup, we can now demonstrate the (mostly) complete
Job-Planning pipeline, starting from a segmentation up to render jobs,
and verify proper connectivity and job invocation.
✔
...ouch this was insidious: the STL implementation for list does not
return a pointer to the element just allocated, but rather retrieves
and dereferences the back() / front() iterator after returning from emplace_back|front()
...which in case of re-entrant allocations is something wildly different
than the initial allocation. Thus a *cheap* and dirty placeholder implementation
just using a STL container is not possible, and we need at least
to code up likewise cheesy placeholder implementation by hand.
- separate allocation and ctor all
- use an inline buffer in the STL container
- explicitly handle ctor failures to discard allocation
- NOT THREADSAFE and likely WASTFUL in terms of performance
==> MockSupport_test now back to GREEN after complete refactoring
...by defining a new scheme for access to custom allocators
...and then passing a reference to such an accessor into the
JobTicket ctor, thereby allowing the ticket istelf recursively
to place further JobTicket instances into the allocation space
--> success, test passes (finally)
...hard to tackle...
The idea is to wrap the TreeExplorer builder, so that our specific
builder functions can delegated to the (inherited) generic builder functions
and would just need to supply some cleverly bound lambdas. However,
resulting types are recursive, which does not play nice with type inference,
and working around that problem leads to capturing a self reference,
which at time of invocation is already invalidated (due to moving the
whole pipeline into the final storage)
...which leads to the next daunting problems:
- we need some mocked ModelPort and DataSink placeholders
- we need a way how to inherit from a partial TreeExplorer pipeline
...introduced in preparation for building the Dispatcher pipeline,
which at its core means to iterate over a sequence of frame positions;
thus we need a way to stop rendering at a predetermined point...
several years ago, it seemed like a good idea to incorporate
the link between nominal time and wall-clock time into a dedicated
anchor point, which also regulates the continued frame planning.
But it turned out that such a design mixes up several concepts
and introduces confusion regarding the meaning of "real time"
- latency can not be reasonably defined for a whole planning chunk
- skipping or sliding due to missed deadlines can not reasonably handled
within such an abstract entity; it must be handled rather at the
level of a playback process
- linking the frame grid generation directly to a planning chunk
undercuts the possible abstraction of a planning pipeline
