Modifications to triangle shader association were not considered when checking for updates and the associated device data array was not tagged as modified so it was not resent to the device(s).
* tbb::blocked_range moved to a different namespace and since the fix is
non-trivial, remove some unused code that used this.
* Task group priorities are no longer supported. It's unclear if they are
useful at all right now and even set correctly, for now all tasks are equal
priority with TBB 2021.
This optimizes device updates (during user edits or frame changes in
the viewport) by avoiding unnecessary computations. To achieve this,
we use a combination of the sockets' update flags as well as some new
flags passed to the various managers when tagging for an update to tell
exactly what the tagging is for (e.g. shader was modified, object was
removed, etc.).
Besides avoiding recomputations, we also avoid resending to the devices
unmodified data arrays, thus reducing bandwidth usage. For OptiX and
Embree, BVH packing was also multithreaded.
The performance improvements may vary depending on the used device (CPU
or GPU), and the content of the scene. Simple scenes (e.g. with no adaptive
subdivision or volumes) rendered using OptiX will benefit from this work
the most.
On average, for a variety of animated scenes, this gives a 3x speedup.
Reviewed By: #cycles, brecht
Maniphest Tasks: T79174
Differential Revision: https://developer.blender.org/D9555
Double floating point precision is an extension of OpenCL, which might
not be implemented by certain drivers, such as Intel Xe graphics.
Cycles does not use double floating point precision, and there is no
need on keeping doubles unless there is an explicit decision to use
them.
This is a simple fix from Cycles side to replace double floating point
type with a type of same size and alignment rules. Inspired by Brecht
and Patrick.
Tested on NVidia Titan V, Radeon RX Vega M, and TGL laptop.
Differential Revision: https://developer.blender.org/D10143
Branched path tracing is not supported for OptiX, and it would still use the
number of AA samples from there when branched path was enabled by the user
earlier but auto disabled and hidden in the UI when using OptiX.
Ref D10159
Tile stealing may steal a CPU tile buffer and move it to the GPU, but next time around that
tile may be re-used on the CPU again (in progressive refinement mode). The buffer would
still be on the GPU then though, so is inaccessible to the CPU. As a result Blender crashed
when the CPU tried to write results to that tile buffer.
This fixes that by ensuring a stolen tile buffer is moved back to the device it is used on before
rendering.
If extensions string is longer than 1024 then the old code would have
reported empty string instead of extensions.
Now the code does dynamic string allocation to store result of request,
similar to what is done in `OpenCLInfo::get_hardware_id`.
The code looks a bit ugly, but it didn't really change much with this
patch. In other words, the code can become more modern and clear, but
it is considered to be outside of the scope of this change.
Differential Revision: https://developer.blender.org/D10135
The OptiX denoiser is part of the OptiX device, so to the tile manager looks like a GPU device. As a
result the tile stealing implementation erroneously stole CPU tiles and moved them to that OptiX
device, even though in this configuration the OptiX device was only set up for denoising and not
rendering. Launching the render kernel therefore caused a crash because of a missing AS etc.
This fixes that by ensuring tiles can only be stolen by devices that support render tiles.
The absolute position desktop mapping has been corrected. The correct
mapping is 0-65535 inclusive. Additionally, division by the virtual
desktop width and height needed to be subtracted by 1 as width and
height are one more than the final pixel index.
Instead of submitting tons of tiny IO syscalls, we can speed things up
significantly by `mmap`ing the .blend file into virtual memory and directly
accessing it.
In my local testing, this speeds up loading the Dweebs file with all its
linked files from 19sec to 10sec (on Linux).
As far as I can see, this should be supported on Linux, OSX and BSD.
For Windows, a second code path uses `CreateFileMapping` and
`MapViewOfFile` to achieve the same result.
Reviewed By: mont29, brecht
Differential Revision: https://developer.blender.org/D8246
When building with more aggressive optimization flags, GCC will add FMA
(Fused Multiply Add) instructions that will slightly alter the floating
point operation results.
This causes some automated tests to fail in blender.
In clang and the intel compiler ffp-contract is set to off per default
it seems from my research. (They do not have the exact same setting,
but the default seems to match the off behavior)
Reviewed By: Brecht
Differential Revision: http://developer.blender.org/D9047
When building with more aggressive optimization flags, GCC will add FMA
(Fused Multiply Add) instructions that will slightly alter the floating
point operation results.
This causes some automated tests to fail in blender.
In clang and the intel compiler ffp-contract is set to off per default
it seems from my research. (They do not have the exact same setting,
but the default seems to match the off behavior)
Reviewed By: Brecht
Differential Revision: http://developer.blender.org/D9047
The absolute position desktop mapping has been corrected. The correct
mapping is 0-65535 inclusive. Additionally, division by the virtual
desktop width and height needed to be subtracted by 1 as width and
height are one more than the final pixel index.
Partially reverts 2250b5cefe.
Removing the user count and fake user count icons was controversial (which was
expected) and there are a few further changes needed, that won't make it in
time for the release, see D9946.
While there is a design to bring back the user count and fake user indicators,
a new design idea was proposed that the UI team wants to follow. This came too
late for the 2.92 release, the new design is targeted at the 2.93 release now.
Meanwhile, UI team decision was to simply revert the design changes.
The new design is being worked on in https://developer.blender.org/T84669.
Note that this commit does not revert some internal changes done in
2250b5cefe. Namely the introduction of `ed_util_ops.c` and data-block
operators in there. These will still be needed in the new design.
In my testing this works, but it requires me to remove the min(start_sample...) part in the
adaptive sampling kernel, and I assume there's a reason why it was there?
Reviewed By: brecht
Maniphest Tasks: T82351
Differential Revision: https://developer.blender.org/D9445
Also send mouse move event instead of moving cursor via SetCursorPos.
SetCursorPos did not reliably move the mouse when it is leaving the
window quickly.
Reviewed By: brecht
Differential Revision: https://developer.blender.org/D9981
The issue is that the "Noisy Image" pass is added even though it should not.
`use_denoising` has to be enabled on the scene and on the view layer
to actually enable it.
Differential Revision: https://developer.blender.org/D10048
Reviewers: lukasstockner97, brecht
Multiple Wintab tablets do not send relative button state when
configured to do so. This causes button events to be delayed until
processed as Win32 button events.
This commit fixes the issue by configuring Wintab to use absolute
button state and tracking changes manually.
Commit d259e7dcfb increased the instance limit, but only provided
a fall back for the host code for older OptiX SDKs, not for kernel code. This caused a mismatch when
an old SDK was used (as is currently the case on buildbot) and subsequent rendering artifacts. This
fixes that by moving the bit that is checked to a common location that works with both old an new
SDK versions.
For a while now OptiX had support for 28-bits of instance IDs, instead of the initial 24-bits (see also
value reported by OPTIX_DEVICE_PROPERTY_LIMIT_MAX_INSTANCE_ID). This change makes use of
that and also adds an error reported when the number of instances an OptiX acceleration structure is
created with goes beyond the limit, to make this clear instead of just rendering an image with artifacts.
Manifest Tasks: T81431
Rendering on the CPU uses the Embree BVH layout, whether the OptiX denoiser is enabled or not.
This means the "build_bvh" function gets a "BVHEmbree" object to fill and not a "BVHMulti" as it
was assuming before, which caused crashes due to memory geting overwritten incorrectly. This
fixes that by redirecting Embree BVH builds to the Embree device.
Manifest Tasks: T83925
Changing the geometry in the current scene caused the primitive offsets for all geometry to
change, but the values would not be updated in all bottom-level BVH structures. Rendering
artifacts and crashes where the result. This fixes that by ensuring all BVH structures are
updated when the primitive offsets change.
Modifies WM_BUTTON processing to reuse existing mousemove logic. Fixes
case where cursor wrap was not being handled on mouse release.
Bonus: flattened mouse move logic so all paths lead to a single return.
Removed case where wrap is not handled until subsequent mousemove as
button press may rely on updated mouse move position.
Previously Wintab packets were added to a local queue to be processed
during Win32 mouse events, in order to correlate Wintab to Win32
mouse buttons. Wintab packets before Win32 mouse down events were
expired on a timer.
This commit drives mouse events during Wintab events when a device is
in range. When a Wintab button is found it is dispatched if an
equivalent event can be popped from the Win32 event queue. If a Win32
mouse button event is not associated with a Wintab event, it falls
through to WM_BUTTON handling. All Wintab packets are handled as they
are received.
Reviewed By: brecht
Differential Revision: https://developer.blender.org/D9908
