| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Use uchar instead of float for vertex data that doesn't need float.
Continue using floats in shaders. Also remove animY2, which is same as
animY1. These changes reduce memory usage especially when the amount
of particles increases.
Testing on windows, memory reductions with emitters/trailemitter
example were:
- OpenGL: 82.7 MB -> 76.5 MB
- Vulkan: 130.8 MB -> 126.3 MB
- D3D11: 143.7 MB -> 135.8 MB
Task-number: QTBUG-88124
Change-Id: I8f8dcb3845323b0e69fb99b5bff830cd0f151a47
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
Reviewed-by: Tomi Korpipää <tomi.korpipaa@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Minor improvements while going through the particles sources:
- Reduce getShadowDatum() calls
- Remove unused m_lastLevel, reloadColor(), m_available
m_velocityFromMovement, maximumAlive, m_pending, m_lastT,
m_last_last_last_emitter, modelIndex
- Color4ub structs can be assigned directly
- Fix pri for shaders_ng files
- Other minor refactoring
Task-number: QTBUG-88124
Change-Id: I5768fed9f4b6379a698849c23436fa38169833a4
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
|
|
|
|
|
|
|
| |
Since we depend on C++17 now, all of these can go.
Change-Id: I0484fd4bb99e4367ec211c29146c316453729959
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
|
|
|
|
|
|
|
|
|
| |
The implementation is intertwined with the direct GL path and not
really portable to RHI.
Fixes: QTBUG-78574
Change-Id: I6ac8707544c19a24661f07e0e6d2fc405e49cbff
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
When an affector acts on a particle, it will add it to a set
of "seen" particles. This means an allocation, per particle,
per frame.
In Qt 6 the problem is less dramatic due to the new QHash
implementation, which uses "wider" buckets, but in Qt 5 this is
hundreds of memory allocations (and deallocations) per frame.
Just reimplement a minimal flat-set API for this use case,
and replace the QSet usages with it. On a testcase with
200 active particles, this reduces memory allocations from
~20'000 per second to 0 when the scene is "stable".
Pick-to: 5.15
Change-Id: I4be1e12a23b8dffca91955148532db243e383a4c
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Reviewed-by: Shawn Rutledge <shawn.rutledge@qt.io>
|
|
|
|
|
| |
Change-Id: Id72fbe10c16de61bd847773d0055d83cfe03f63c
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
|
|\
| |
| |
| |
| |
| |
| | |
Conflicts:
src/qml/common/qv4compileddata_p.h
Change-Id: I1150c8cd0161f0e22137d383013751394ae64e18
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
The infinite loop was triggered by several issues coincide together.
In short, the direct cause is the particle's born time and lifespan were
represented in 32 bit floats and not precise enough to pass aliveness check as
time grows large.
While the time grows large, the resolution of floating point decreases to the
extent that resolution is even bigger than 2 milliseconds.
Then it will fail to pass the aliveness check. Then, the
dead particles will be treated alive and they are kept inserting into and
popping out of the particles heap, which is similar to a live-lock.
The fix is to separate freeing dead and inserting back alive ones in two
different loops, ensure that the emitter can update time for next frame.
There are still other issues:
1) as the times runs very long, the particle needs several frames's updates
to actually make the states change noticeable, which means animation may
become not so smooth after running for too long (like several days).
May change particle's born/lifespan time to 64 bits in another patch.
2) the particle system's and animation's timers are 32 bit integers,
after 2^31 milliseconds(24.8551348 days), they will overflow. May promote
them to 64 bits in another patch.
3) as the time grows even larger such that the resolution is bigger than 16ms
at 60 hz frame rate, the live-lock may occur again. Because the timer advances/delta
will be not large enough to make dead ones reused.
The next live-lock estimated time is 2^24*16 milliseconds = 3.10689185 days.
The final fixes are 1) and 2)
4) may change the particle system's internal timer be set to arbitrary value
(fast forward to large value) for easier writing autotest for above cases.
Change-Id: I1190c0814c8197876b26dd4182dc4b065dd1ece6
Task-number: QTBUG-64138
Reviewed-by: Vitaly Fanaskov <vitaly.fanaskov@qt.io>
|
|/
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Using this technique we can automatically register all necessary
revisions and minor versions of a type, using the metaobject system.
This greatly reduces the potential for mistakes and resulting
incompatibilities between versions of imports.
We assume that for each type we need to register all revisions of its
super types and its attached type, and that the revisions match. That
is, if you import version X of type A, you will also get version X of
its attached type and of any super types. As we previously didn't take
these dependencies into account when manually registering the types, a
number of extra revisions are now registered for some types.
Potentially, we can now generate the qmltypes files at compile time,
using moc.
Change-Id: I7abb8a5c39f5e63ad1a0cb41a783f2c91909491b
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
|
|
|
|
|
|
|
|
| |
Provide different export macros and different top level headers for
each, don't include runtime headers from compiler sources.
Change-Id: I7dc3f8c95839a00a871ba045ec65af87123154be
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
|
|
|
|
|
|
|
|
| |
Move the relevant files into more fitting locations and build the
devtools from only parser, compiler and qmldirparser.
Change-Id: Ibf37a1187f36d02983f9f43c6622acb243785b7b
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
|
|
|
|
|
|
|
|
| |
This is just an alias for QV4::ReturnedValue. We can as well use the
latter.
Change-Id: Ibd2c038a3ca726b39a8f0f05e02922adb9fccbdb
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
|
|
|
|
|
|
|
|
| |
clang-tidy -p compile_commands.json $file -checks='-*,modernize-use-default-member-init,readability-redundant-member-init'
-config='{CheckOptions: [{key: modernize-use-default-member-init.UseAssignment, value: "1"}]}' -header-filter='qtdeclarative' -fix
Change-Id: I705f3235ff129ba68b0d8dad54a083e29fcead5f
Reviewed-by: Johan Helsing <johan.helsing@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
From now on we prefer nullptr instead of 0 to clarify cases where
we are assigning or testing a pointer rather than a numeric zero.
Also, replaced cases where 0 was passed as Qt::KeyboardModifiers
with Qt::NoModifier (clang-tidy replaced them with nullptr, which
waas wrong, so it was just as well to make the tests more readable
rather than to revert those lines).
Change-Id: I4735d35e4d9f42db5216862ce091429eadc6e65d
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
QSignalMapper ought to be deprecated soon.
It simplifies the code, too.
There is still one use in QQuickGenericShaderEffect which is a bit complicated
to get rid of. A very similar use of QSignalMapper was in use in
QQuickOpenGLShaderEffectCommon but was removed in commit 8c745d80, the same
should be done for QQuickGenericShaderEffect.
(Note the QueuedConnection in qquickparticlesystem is there because the
QSignalMapper used to be in the main thread, meaning a round-trip via the
event loop)
Change-Id: I331b787becbad37f717035bf119bafd7a7214630
Reviewed-by: Frederik Gladhorn <frederik.gladhorn@qt.io>
|
|
|
|
|
|
|
| |
... and drop redundant 'virtual'
Change-Id: Id70a0b74724d231bae9b3803fcfe0514adb9f642
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
When a particle emitter is not yet complete, calling emittersChanged
through both groupChanged and systemChanged is unnecessary. Only connect
to those signals once after the component is complete, and then call
emittersChanged.
Also do not call emittersChanged when the particle system component is
not complete yet.
Change-Id: I67543ce80c9235bd5f72d95352ec96e97a1cf66b
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
|
|
|
|
|
|
|
|
|
|
|
|
| |
According to IEEE754 a * t * t is not the same as a * (t * t), hence the
compiler won't lift t^2 out as common expression when in another place
'a' is replaced with 'b'. Now for the particles this doesn't matter as
it doesn't need super high precision, nor should any of the calculations
come in overflow territory. So by manually introducing t_sq(uare), the
compiler can remove duplicate t^2 calculations after inlining.
Change-Id: Ibfdb054e34945d7e78eb993fdcedfae886472e83
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
|
|
|
|
|
|
|
|
| |
This removes float<->double conversions, which makes the auto-vectorized
code even smaller.
Change-Id: Ic4319b11a3248e3034b65e7be047d99cba84716b
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
|
|
|
|
|
|
|
|
|
|
| |
Calls to these functions often occur in sets of two (x, y) or even more
(x, y, vx, vy, etc). By allowing the compiler to inline, it allows for
many CSE opportunities. Also, if your compiler is reasonably good, it
will also auto-vectorize the operations.
Change-Id: I4bffe4826671dd60683b941a569fc6a7b4b34da7
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The reusableIndexes represented a "free-list". Now the allocation
behavior in QQuickParticleGroupData::setSize was to grow by (large)
chunks. That means that as soon setSize was called, a (big) number of
hash entries was created, which are drained over time. This memory would
stay around (and probably unused) as long as the group was alive.
By using a bit vector, the amount of memory is much more compressed,
and finding an entry takes less time. The FreeList "caches" the next
free entry, because allocation and de-allocation behavior is that they
occur bunches: allocate a number of particles, use them, allocate the
same number.
Test case: samegame, 1 player, click 1 set of 3 stones, quit.
QQuickParticleSystem::emittersChanged(), before patch:
- 21 instr. inclusive, 15M in QQuickParticleGroupData::setSize
- 23,000 calls to QHashData::allocateNode
after:
- 13M instr. inclusive, 7M in QQuickParticleGroupData::setSize
- 0 calls to QHashData::allocateNode
Change-Id: If35ea5ed9b29129f210638f6f59275a24eb6afdc
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
|
|
|
|
|
|
|
|
| |
Samegame creates about 23,000 particles, so this reduces the memory by
~180kb on 32bit, so ~360kb on 64bit.
Change-Id: I0581524ab232b474c5d43abeabd7ebf6174e740f
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
By caching the group id in the particle emitter, and groups in the
painter.
Test case: samegame, 1 player, click 1 set of 3 stones, quit.
QQuickParticleSystem::emittersChanged(), before patch:
- 64M instr. inclusive
- 145,880 calls to findNode (29M instr.)
after:
- 21M instr. inclusive
- 0 calls to findNode
- 25 calls to QQuickParticlePainter::recalculateGroupIds (9800 instr.
inclusive).
Change-Id: I4aba9d50100513c6b7cdd230e30b3aecaf84485a
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
|
|
|
|
|
|
|
|
|
|
|
| |
And when in the area, remove 3 unused functions.
Benchmark: start samegame, click '1 player', quit.
Valgrind result: removes/inlines ~90,000 calls.
Change-Id: If911b19b46bf163f7fe678623c068f960296f17e
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
Reviewed-by: Robin Burchell <robin.burchell@viroteck.net>
|
|
|
|
|
|
|
|
|
|
|
| |
Replace QHash<int, QQuickParticleGroupData*> groupData with a
QVarLengthArray, and make sure that those integers (== indices) are
continuous, re-used, and start at zero. That way a whole bunch of qhash
calls, hash node creatrion/deletion, and other overhead is removed.
Change-Id: Ie74fab8a3e3c7b6efa15b7b9ceff1d1a3e9820e9
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
Reviewed-by: Robin Burchell <robin.burchell@viroteck.net>
|
|
|
|
|
|
|
|
|
|
|
| |
From Qt 5.7 -> LGPL v2.1 isn't an option anymore, see
http://blog.qt.io/blog/2016/01/13/new-agreement-with-the-kde-free-qt-foundation/
Updated license headers to use new LGPL header instead of LGPL21 one
(in those files which will be under LGPL v3)
Change-Id: Ic36f1a0a1436fe6ac6eeca8c2375a79857e9cb12
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
There are several special cases where the particle system is very
useful for generating more than just eye candy. An example is a
recreation of world weather/wind conditions as seen in
http://earth.nullschool.net/ this patch allows for the implementation
of similar systems.
This patch does not adds features, just adds the macro
Q_QUICK_PRIVATE_EXPORT in order to link the classes and derive
custom C++ affectors for Qt Quick usage.
Task-number: QTBUG-41502
Change-Id: I809d74676b3be73221840c694d7ca3eebe553c02
Reviewed-by: Gunnar Sletta <gunnar@sletta.org>
|
|
|
|
|
|
| |
Task-number: QTBUG-48594
Change-Id: Ifc207938de7f0c8995fc712df92665f222612647
Reviewed-by: Alan Alpert <aalpert@blackberry.com>
|
|
|
|
|
|
|
| |
x86_64.
Change-Id: Ide7a70baecf6d1ff3ec8160a2cb47f0bd9955e45
Reviewed-by: Alan Alpert <aalpert@blackberry.com>
|
|
|
|
|
|
|
|
|
| |
Qt copyrights are now in The Qt Company, so we could update the source
code headers accordingly. In the same go we should also fix the links to
point to qt.io.
Change-Id: I61120571787870c0ed17066afb31779b1e6e30e9
Reviewed-by: Iikka Eklund <iikka.eklund@theqtcompany.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Make sure that ImageData instances, m_shadowData as well as the
QSGGeometry of particle nodes are destroyed together with their
QQuickImageParticle.
Also implement the assignment operator for QQuickParticleData to
avoid its v8Datum pointer to be copied over to the shadow datum
in getShadowDatum. This would cause a double delete of the
QQuickV4ParticleData when trying to call clearShadows() in the
destructor.
Task-number: QTBUG-36782
Change-Id: Ie03f2be0415daeb7f4f6e5f92295a3ab26a62155
Reviewed-by: Joerg Bornemann <joerg.bornemann@theqtcompany.com>
|
|
|
|
|
|
|
|
|
| |
- Renamed LICENSE.LGPL to LICENSE.LGPLv21
- Added LICENSE.LGPLv3 & LICENSE.GPLv2
- Removed LICENSE.GPL
Change-Id: I84a565e2e0caa3b76bf291a7d188a57a4b00e1b0
Reviewed-by: Jani Heikkinen <jani.heikkinen@digia.com>
|
|
|
|
|
|
|
|
|
|
|
| |
This is necessary for the QtWebEngine module.
This also adds an empty nokeywords test using the same mechanism
as qtbase/tests/auto/tools/moc/no-keywords.h to find conflicts
at compile time.
Change-Id: I9df541720797dd61f078178c2af68ead18ff8bfe
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
|
|
|
|
|
| |
Change-Id: Ifc48b73a2fbfda22d3c4419cb34dd5307107f688
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
|
|
|
|
|
| |
Change-Id: I87d2183738ec7cfeea846a28f2b9aed79a233f68
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
|
|
|
|
|
|
|
|
|
| |
The handle wraps a V4 Value, so this is the better name for it.
Also added some accessor methods to convert to and from V4
Values.
Change-Id: I327c83feb5bd3be59909001489979e5a3a9d9e67
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
|
|
|
|
|
|
|
|
|
|
|
| |
The macro was made empty in qtbase/ba3dc5f3b56d1fab6fe37fe7ae08096d7dc68bcb
and is no longer necessary or used.
Discussed-on: http://lists.qt-project.org/pipermail/development/2013-January/009284.html
Change-Id: Ia07e99676e0134fde5e32880edb95e57c779a7ff
Reviewed-by: Laszlo Papp <lpapp@kde.org>
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@digia.com>
Reviewed-by: Alan Alpert <aalpert@rim.com>
|
|
|
|
|
|
| |
Change-Id: I6c3bd7bebe3d62d1cfd0fa6334544c9db8398c76
Reviewed-by: Akseli Salovaara <akseli.salovaara@digia.com>
Reviewed-by: Sergio Ahumada <sergio.ahumada@digia.com>
|
|
|
|
|
|
|
| |
Change copyrights and license headers from Nokia to Digia
Change-Id: Ie7f5d49ed8235d7a7845ab68f99ad1c220e64d5c
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
|
|
|
|
|
|
|
|
| |
Matches the convention set in the QtQuick module, for example by
ListView and Flickable.
Change-Id: I8df57ed1ced8128723d790c30c00cc1b2062787d
Reviewed-by: Bea Lam <bea.lam@nokia.com>
|
|
|
|
|
| |
Change-Id: Ibf14dc0f9ef9e17225f07f31276085c9ca83d4f1
Reviewed-by: Yunqiao Yin <charles.yin@nokia.com>
|
|
QtQuick clients that do not use particles features should not have
to load these classes.
Task-number: QTBUG-25178
Change-Id: Ib15f7655dc4d821595e06f9160d2770375279027
Reviewed-by: Glenn Watson <glenn.watson@nokia.com>
|