Commit Graph
6478 Commits
Author SHA1 Message Date
Powei Feng cdd0147a17 Release Filament 1.53.5 2024-08-13 15:59:08 -07:00
Sungun Park 9de29a475e Add warning for the incompatibility of stereo type (#8031)
Print a warning in case the sterescopic type in a compiled package is
different than what's in the engine's setting. The application may
proceed, but it could end up visual glitches when enabling stereoscopic
rendering.

This requires the stereoscopic type to be written into the package,
which needs a material version bump.
2024-08-13 01:31:10 +00:00
Mathias Agopian 7f97363d9c fix debug build 2024-08-12 15:09:06 -07:00
Mathias Agopian d88b4c1bc3 Add a way to retrieve the count of all resources
This is intended for debugging.

BUGS=[358111049]
2024-08-12 14:50:43 -07:00
Mathias Agopian 3badf9db18 read all valid pending timer queries each frame
This reduces the latency of the timer query result; with the previous
code the latency could only increase, but there is no reason to wait
a whole frame for reading the next available result.

We just loop over them until we find one that has not signaled; instead
of doing one per frame.
2024-08-12 14:50:20 -07:00
Mathias Agopian 929f793cf3 clean-up RenderPass a bit
make mCommandBegin and mCommandEnd const* const, this allows to make
sort() And resize() static, and instanceify() almost static..
2024-08-12 14:49:57 -07:00
Mathias Agopian 06bffaa650 fix a couple precondition checks logs 2024-08-12 14:49:44 -07:00
Ben Doherty 2e2f111435 Re-enable missing packing/unpacking functions in ESSL, fix IMG shader compilation (#8030) 2024-08-12 10:23:52 -07:00
Mathias Agopian 06f9626429 improve GL backend debug markers
we now have two levels of debug markers. Those that come from the "user"
(i.e. filament itself) are now always enabled and generate both 
systrace and gl markers. the 2nd level is internal and always
disabled by default. Of enabled at compile time it'll emit markers for
each driver API method.
2024-08-09 15:25:33 -07:00
Mathias Agopian fa0e20a699 always validate Engine::config 2024-08-08 22:46:57 -07:00
Mathias Agopian 59593830e5 Fix skipFrame incorrectly asserts in some cases
BUGS=[357992376]
2024-08-08 13:24:38 -07:00
Mathias Agopian 112f9d742c assert when a program compilation fails
The current behavior is to get a UB, which usually is a crash. So this
is better.
2024-08-08 13:24:08 -07:00
Ben Doherty 8ba20eb03c Metal: fix static texture target on more devices (#8022) 2024-08-08 13:04:42 -07:00
Ben Doherty 1c0370d5d7 Remove flaky assertion in ShadowMap (#8023) 2024-08-08 13:04:30 -07:00
Benjamin Doherty 56f714633d Release Filament 1.53.4 2024-08-08 12:12:27 -07:00
Benjamin Doherty f0bc338c80 Revert: inject the missing packing/unpacking function in ESSL 3.0 2024-08-07 16:18:02 -07:00
Benjamin Doherty 9750ddb9db Add type_traits header 2024-08-05 10:59:46 -07:00
Mathias Agopian ec2ee9db7a minor cleanups in CommanmdBufferQueue
- use the same code on both ends of updating the free space. i.e.
  both side compute the "used" space in exactly the same way.
  the math was the same before, but the code was different which
  could be confusing.

- assert for overflow before queuing the buffer. It wouldn't matter
  anyways, because it's done with the condition lock held, so the
  consumer would never have a chance to deuque it. still, less 
  confusing.
2024-08-05 10:06:44 -07:00
Mathias Agopian 1ce1e335a0 gl backend: add a couple missing systrace events 2024-08-02 23:21:34 -07:00
Sungun Park 5fe356e446 Fix ResourceAllocator compile error on Windows (#8011)
This commit 730bc99025 introduced a new
dependency on ResourceAllocator because of the new field
`std::unique_ptr<ResourceAllocator> mResourceAllocator{};` in
details/Renderer.h

This requires cpp files including details/Renderer.h to include
ResourceAllocator.h as well.

This compile issue only happens on the Windows compiler, Visual Studio.
2024-08-02 14:54:02 -07:00
Eliza Velasquez 14ddadd2f6 Fix ES2/anisotropy bug in updateSamplerGroup
This function attempts to set texture parameters in these two cases, but the
texture is not guaranteed to be bound. Perhaps it once was, but the assumption
broke at some point.
2024-08-01 16:44:52 -07:00
Mathias Agopian a2f5b635d1 don't mix seq_cst with other memory orders
this change shouldn't have any impact on ARM, however, according
to cppreference it's not safe to mix seq_cst with other memory
orders:

"as soon as atomic operations that are not tagged memory_order_seq_cst 
enter the picture, the sequential consistency guarantee for the program 
is lost"
2024-07-31 10:11:01 -07:00
Mathias Agopian 324dcd3c86 early exit in fog to improve performance 2024-07-31 10:10:48 -07:00
Sungun Park 4125802644 Release Filament 1.53.3 2024-07-31 16:22:54 +00:00
Balaji M 2d21fcbe55 [gltfio] Initialize mCgltfBuffersLoaded to fix crash (#7999)
Value of mCgltfBuffersLoaded is sometimes retained across creation of FAssetLoader which skips loading the buffer in AssetLoaderExtended#createPrimitive leading to null pointer crash
2024-07-31 02:12:04 +00:00
Sungun Park 3f4dac6248 Fix a crash for IBL resource loading (#8001)
This fixes a crash introduced by a8ace2891d

The refactored FrameInfoManager can cause a crash when IBL resource loading
happens because now the getLastFrameInfo() references an invalid value via the
`front` method. Return the default FrameInfo to resolve this.

Also fix a null pointer reference bug for OpenGLTimer::State, which
happenes when the renderer for IBLPrefilterContext is destroyed.
2024-07-29 22:25:36 +00:00
Ben Doherty 40ae416715 Metal: redistribute pool sizes (#8000) 2024-07-29 13:14:50 -06:00
Grant CommodoreandPowei Feng b54a2046cb Multiview support for vulkan (#7892)
Adds multiview support for vulkan. This is done by adding a layerCount to the renderTarget, which is used to determine if multiview is available and being used in the current renderpass.

FIXES=[332425392]

Co-authored-by: Powei Feng <powei@google.com>
2024-07-26 07:41:00 +00:00
Mathias Agopian 8add6ae1ac fix a potential crash is AssetLoader
if it can't find a name for a node, it will revert to the config's
defaultNodeName, however, if that is nullptr also, a crash will occur.
so we provide a last-resort hardcoded name in that case.
2024-07-25 22:44:34 -07:00
Mathias Agopian e7c96cd124 better handle skipped frames and cache eviction
Add Renderer::skipFrame() which should be called when intentionally 
skipping frames, for instance because the screen content hasn't changed,
allowing Filament to performance needed periodic tasks, such as
cache garbage collection and callback dispatches.

We also improve the ResourceAllocator cache eviction policy:
- the cache is aggressively purged when skipping a frame
- we aggressively evict entries older than 3 frames

The default Config is now set to a more agressive setting.
2024-07-25 17:23:04 -07:00
Mathias Agopian 730bc99025 Move the ResourceAllocator from Engine to Renderer
The ResourceAllocator used to be global and owned by the Engine, this
was causing some issues when using several Renderers because each
one could cause the eviction of cache data for another.

We now have a ResourceAllocator per Renderer, which makes more sense
because most resources are allocated by the FrameGraph.

We also introduce a ResourceAllocatorDisposer class, which is used
for checking in and out a texture from the cache, and destroy the
texture when it's checked-out. That objet is still global.
2024-07-25 17:23:04 -07:00
Powei Feng 7441e878bb gltfio: enable escaped unicode for node name (#7989)
Fixes #7846
2024-07-25 09:06:56 +00:00
Grant CommodoreandPowei Feng ef4c4a76fc Only sets the layercount if the view has stereo and the stereoscopic type is multiview (#7987)
Co-authored-by: Powei Feng <powei@google.com>
2024-07-25 07:31:51 +00:00
Mathias Agopian be4f287b07 More improvements to the JobSystem (#7988)
* improve parallel_for a bit

We get about 40% performance increase. The gain comes from not having
to copy the JobData structure each time we create a job, by using a
new emplaceJob() method, we can create the structure directly into
its destination.

* avoid calling wakeAll() when possible

wakeAll() is very expensive and not always needed when a job finishes
because there may not be anyone waiting on that job.

We now maintain a waiter count per job, and use that to determine if
we need to notify or not.

And now that the JobSystem overhead is lower, we can decrease the size
of the jobs, which improves the load balancing.

* mActiveJobs fixes

some comments claimed mActiveJobs needed to be modified before or after
accessing the WorkQueue; this couldn't be correct because there were no
guaranteed global ordering with the workQueue.
2024-07-24 15:34:36 -07:00
Mathias Agopian d3a35de386 fix a crash with bloom when screen dimension smaller than 16px 2024-07-24 11:49:38 -07:00
Sungun Park 9057c47a56 Release Filament 1.53.2 2024-07-23 01:23:57 +00:00
Mathias Agopian 4a5081388d RenderPass batch size was too small
The batch size was calculated incorrectly, which in theory could lead
to command buffer overflows (albeit unlikely).
2024-07-22 13:53:05 -07:00
Mathias Agopian a60fe41681 performance improvements to JobSystem
- reduce the number of calls to notify_one() and notify_all().
  notify_one() is not only called when running a new job, and
  notify_all() only when a job finishes.

- don't hold the condition lock while calling notify_*(), as it is not
  strictly needed, and because notify_*() can be very slow, there can
  be a lot of contention on this lock as a result; blocking the whole
  jobsystem thread pool.

- add a new version of run() that takes an opaque thread id that can
  be retrieved from a job's execute function; this is especially
  intended to be used by parallel_for(); it's just a more efficient
  version of run() that avoids a hashmap lookup.


Overall these change yield a significant performance boost:
- running + waiting a job: +200%
- running many jobs: +150%
- running many jobs in parallel: +50%
2024-07-22 11:12:51 -07:00
Mathias Agopian a8ace2891d improve frame timing info
- we use a circular buffer for the frame history so that 
  we don't have to copy the data when insert a new entry.
  This also allows us to keep a reference to an entry, which
  doesn't get invalidated when an entry is added/removed.

- we now store the gpu frame time in the correct slot (instead of
  always the latest). It didn't matter before because the API wasn't
  public and we only needed some recent frame time.

- a new public API now returns the frame history, which now contains 
  more data; in particular the main and backend thread's begin/end
  frame time.


BUGS=[321110544]
2024-07-18 21:14:44 -07:00
mdagoisandPowei Feng 669ffc85c0 Minor typo fix and slight optimization (#7978)
Co-authored-by: Powei Feng <powei@google.com>
2024-07-19 03:07:54 +00:00
mdagoisandPowei Feng d957dd8082 Fix to barriers in VulkanBuffer (#7979)
Co-authored-by: Powei Feng <powei@google.com>
2024-07-19 10:45:10 +08:00
Ben Doherty 8eda62c957 Update imgui to v1.90.9 (#7977) 2024-07-18 16:38:25 -07:00
Sungun Park 90f079ac2c Add an extra check for renderStandaloneView (#7975)
renderStandaloneView must be called outside of beginFrame() /
endFrame(). This extra check ensures this behavior.
2024-07-18 06:31:02 -07:00
Balaji M 3728f06603 utility::loadCgltfBuffers is done once instead of doing for each primitive (#7969) 2024-07-12 23:22:25 +00:00
Adam Wychowaniec 39cfce641c Fix precondition check when creating static geometry. 2024-07-08 11:53:24 -07:00
Mathias Agopian 3fbf3f7111 fix a couple small bugs when mapping the circular buffer
- in case of failure we were munmap'ing the wrong size for the
  guard page (in practice this never happened)
- the post-condition check was incorrect; it checked for nullptr instead
  of MAP_FAILED. this also never happened in practice.

Also made a couple of small improvements:

- in case the special circular buffer mapping fails, log a message
  as warning instead of debug.

- immediately memset (i.e. populate) the pages for the circular buffer
  since they will all be accessed rather quickly.
2024-07-08 10:48:44 -07:00
Reinar ef9bbece2a Fix feature level detection for VulkanDriver 2024-07-02 15:09:25 -07:00
Mathias Agopian eb8fa9ecdf shader compilation could fail on ES 3.0 devices
the failure could happen if the shader didn't have any #extension
strings, which was likely to happen on release builds (e.g. on 
emulator).
2024-07-01 11:25:12 -07:00
Ben Doherty e1bfe7ca81 Metal: add a more detailed CHECK_POSTCONDITION when creating a texture (#7943) 2024-06-28 13:11:00 -07:00
Ben Doherty ca4c7ac739 Update cgltf to 1.14 (#7945) 2024-06-28 13:10:41 -07:00