This patch replaces `dispatchThreadgroups` with `dispatchThreads` which takes care of non-uniform threadgroup bounds. This allows us to remove the bounds guards in the integrator kernel entry points.
Pull Request: https://projects.blender.org/blender/blender/pulls/106217
For example
```
OIIOOutputDriver::~OIIOOutputDriver()
{
}
```
becomes
```
OIIOOutputDriver::~OIIOOutputDriver() {}
```
Saves quite some vertical space, which is especially handy for
constructors.
Pull Request: https://projects.blender.org/blender/blender/pulls/105594
Workaround for a crash when `addComputePipelineFunctionsWithDescriptor` is called *after* `newComputePipelineStateWithDescriptor` with linked functions (i.e. with MetalRT enabled). Ideally we would like to call `newComputePipelineStateWithDescriptor` (async) first so we can bail out if needed, but we can stop the crash by flipping the order when there are linked functions. However when addComputePipelineFunctionsWithDescriptor is called first it will block while it builds the pipeline, offering no way of bailing out.
Note that this only has an impact when the "MetalRT (Experimental)" option is checked.
Pull Request: https://projects.blender.org/blender/blender/pulls/105629
The traversable handle of a BLAS may be zero when the relevant geometry
is empty (no triangles/curves/points/...), as no BLAS is built in such cases.
It is not correct to attach a zero handle to a TLAS, so filter out such instances.
When running unit tests or other fast completing renders, forced crashes can occur if there are any slow, outstanding PSO compilation requests (due to the `std::terminate` fall-back case in `~ShaderCache`).
This patch eliminates the need for this shutdown hack by using of the async version of `newComputePipelineStateWithDescriptor` when creating a PSO for the first time. In doing so, we are able to explicitly respond to app shutdown instead of waiting for the pipeline to finish compiling (..and then timing out and force-crashing). We still use the blocking version of `newComputePipelineStateWithDescriptor` when loading from an archive, as this can handle loading from a corrupted archive gracefully. Finally, we move `addComputePipelineFunctionsWithDescriptor` to *after* the PSO is built (as this will trigger a full blocking compile if the PSO has not yet been built, which would bring back the original issue).
Pull Request: https://projects.blender.org/blender/blender/pulls/105506
This patch fixes hanging unit tests when MetalRT is enabled. It simplifies and fixes the kernel selection logic by baking the MetalRT-specific options into `kernels_md5` rather than expanding out and testing MetalRT bit flags explicitly.
Pull Request #105270
To improve mesh upload speeds and reduce the size of the scene data which allows larger scenes to be rendered.
The meshes in Cycles are currently stored as flattened meshes, where each triangle is stored as a set of 3 vertices. Unflattening writes out the vertices in a list according to the index buffer. This uses a lot of memory and for current hardware does not provide a noticeable benefit. This change unflattens the mesh by directly using the meshes vertex and index buffers directly and skips the unflattening. This change allows for larger scenes and also a reduction in the sizes of the meshes. Further it results in a decrease the amount of time it takes to upload the data to a GPU. This is especially important for when multiple GPUs are used in a single machine.
Pull Request #105173
This is a workaround for [issue #104087](https://projects.blender.org/blender/blender/issues/104087). We encounter crashes when using shader binary archives on AMD, so this disables them while we investigate a proper fix. Kernels will still be cached automatically by the OS file system cache. This cache may occasionally be purged due to external factors, in which case kernels will get compiled again.
Pull Request #105186
This fixes issue [#105100](https://projects.blender.org/blender/blender/issues/105100) where multi-pass renders can be incorrect due to kernels using stale specialisation constants (e.g. when rendering Pokedstudio).
This patch adds a new group of md5 hashes (`global_defines_md5`) to track whether the injected block of #defines is stale and regenerate the source string as appropriate. It also renames the existing group of md5 hashes from `source_md5` to `kernels_md5` to clarify that these refer to a specific kernel set rather than just the source (which might build an arbitrarily large number of kernel sets).
Pull Request #105103
Contributed by Yulia Kuznetcova at Apple.
NanoVDB is patched to give add address spaces required by Metal. We hope that
in the future Metal will support the generic address space.
For AMD and Intel this is currently not available since it causes a performance
regression also on scenes without volumes.
Pull Request #104837
The internal compiler error appears to be gone. Unclear why it appeared in the
first place and why it's gone now. Just random kernel code changes causing it.
Pull Request #104719
(Follow on from D17043)
On AMD Navi2 devices the MetalRT checkbox was not hooked up properly and had no effect. This patch fixes it.
Co-authored-by: Michael Jones <michael_p_jones@apple.com>
Pull Request #104520
Host memory fallback in CUDA and HIP devices is almost identical.
We remove duplicated code and create a shared generic version that
other devices (oneAPI) will be able to use.
Reviewed By: brecht
Differential Revision: https://developer.blender.org/D17173
This patch optimises subsurface intersection queries on MetalRT. Currently intersect_local traverses from the scene root, retrospectively discarding all non-local hits. Using a lookup of bottom level acceleration structures, we can explicitly query only the relevant instance. On M1 Max, with MetalRT selected, this can give a render speedup of 15-20% for scenes like Monster which make heavy use of subsurface scattering.
Patch authored by Marco Giordano.
Reviewed By: brecht
Differential Revision: https://developer.blender.org/D17153
This patch adds two new kernels: SORT_BUCKET_PASS and SORT_WRITE_PASS. These replace PREFIX_SUM and SORTED_PATHS_ARRAY on supported devices (currently implemented on Metal, but will be trivial to enable on the other backends). The new kernels exploit sort partitioning (see D15331) by sorting each partition separately using local atomics. This can give an overall render speedup of 2-3% depending on architecture. As before, we fall back to the original non-partitioned sorting when the shader count is "too high".
Reviewed By: brecht
Differential Revision: https://developer.blender.org/D16909
This patch removes the option to select both AMD and Intel GPUs on system that have both. Currently both devices will be selected by default which results in crashes and other poorly understood behaviour. This patch adds precedence for using any discrete AMD GPU over an integrated Intel one. This can be overridden with CYCLES_METAL_FORCE_INTEL.
Reviewed By: brecht
Differential Revision: https://developer.blender.org/D17166
This patch fixes T103393 by undefining `__LIGHT_TREE__` on Metal/AMD as it has an unexpected & major impact on performance even when light trees are not in use.
Patch authored by Prakash Kamliya.
Reviewed By: brecht
Maniphest Tasks: T103393
Differential Revision: https://developer.blender.org/D17167
A noticeable (>5%) performance regression in oneAPI backend came with
a501a2dbff. Updating to latest graphics
compiler from driver 101.4032 fixes it.
I've tested it with current min-supported drivers and it runs well but
since compatibility of graphics compiler with older drivers isn't
guaranteed, I'm also bumping the min-supported driver versions.
If end-users consider latest drivers too fresh to switch to (version
isn't released as stable on Linux as of today but should be before
Blender 3.5 release), CYCLES_ONEAPI_ALL_DEVICES=1 env variable can be
used.
Intel Graphics Compiler on Linux will be updated in a later commit
so we can then close D16984.
Reviewed By: sergey, LazyDodo
This patch adds markup to specify that certain kernel data constants should not be specialised. Currently it is used for `tabulated_sobol_sequence_size` and `sobol_index_mask` which change frequently based on the aa sample count, trash the shader cache, and have little bearing on performance.
Reviewed By: brecht
Differential Revision: https://developer.blender.org/D16968
This patch adds occupancy tuning for the newly announced high-end M2 machines, giving 10-15% render speedup over a pre-tuned build.
Reviewed By: brecht
Differential Revision: https://developer.blender.org/D17037
While keeping SSE2, SSE4.1 and AVX2. This does not affect hardware support, it
only slightly reduces performance for some older CPUs.
To reduce maintenance cost and improve compile times.
Differential Revision: https://developer.blender.org/D16978
