OpenImageDenoise V2 comes with GPU support for various backends. This adds a new class, OIDNDenoiserGPU, in order to add this functionality into the existing Cycles post processing pipeline without having to change it much. OptiX and OIDN CPU denoising remain as they are. Rendering on a supported Intel GPU will automatically select the GPU denoiser.
Device support is initially limited to the oneAPI devices that are supported by Cycles, but can be extended.
Ref #115045
Co-authored-by: Stefan Werner <stefan.werner@intel.com>
Co-authored-by: Ray Molenkamp <github@lazydodo.com>
Pull Request: https://projects.blender.org/blender/blender/pulls/108314
In order to speedup compilation, we upgrade IGC to 1.0.14828.26 along
with ocloc and the associated dependencies.
We also bump min-driver version accordingly to 26918.
Ref !114341
The first public Windows driver version with a higher number is
101.4824, so we bump the min-required driver version on Windows to this
one to ensure compatibility.
<algorithm> header include is missing from some sycl headers, this will
be fixed upstream with https://github.com/intel/llvm/pull/10424,
meanwhile, we work around it by including it directly.
Recent versions of DPC++ dropped using the environment variable
SYCL_PI_LEVEL_ZERO_USE_COPY_ENGINE_FOR_IN_ORDER_QUEUE=0 we were setting.
We're now also setting SYCL_PI_LEVEL_ZERO_USE_COPY_ENGINE=0 by default
to keep a consistent behavior.
During recent testing, the oldest 101.4032 (windows) and <25812 (linux)
drivers led to crashes during JIT compilation, so we bump the
requirement to newer 101.4313 and 25812.14 drivers that do incorporate
the required fixes.
Pull Request: https://projects.blender.org/blender/blender/pulls/109281
Only Embree CPU BVH was built in the multi-device case. However, one
Embree GPU BVH is needed per GPU, so we now reuse the same logic as in
the other backends.
Pull Request: https://projects.blender.org/blender/blender/pulls/107992
Updated Embree 4 library with GPU support is required for it to be
compiled - compatiblity with Embree 3 and Embree 4 without GPU support
is maintained.
Enabling hardware raytracing is an opt-in user setting for now.
Pull Request: https://projects.blender.org/blender/blender/pulls/106266
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 functionality is related only to debugging of SYCL implementation
via single-threaded CPU execution and is disabled by default.
Host device has been deprecated in SYCL 2020 spec and we removed it
in 305b92e05f.
Since this is still very useful for debugging, we're restoring a
similar functionality here through SYCL 2020 Host Task.
This patch tunes the integrator state sizing for Metal (`num_concurrent_states` and `num_concurrent_busy_states`).
On all GPUs architecture, we adjust the busy:total states ratio to be 1:4 which gives better rendering performance than the previous 1:16 ratio (independent of total state count). This gives a small performance uplift (e.g. 2-3% on M1 Ultra).
Additionally for M2 architectures, we double the overall state size if there is available headroom. Inclusive of the first change, we can expect uplift of close to 10% in future, as this results in larger dispatch sizes and minimises work submission overheads. In order to make an accurate determination of available headroom, we defer the calculation of `num_concurrent_states` and `num_concurrent_busy_states` until the time of integrator state allocation (i.e. after all of the scene data has been allocated). We also refactor `alloc_integrator_soa` to calculate an *exact* single-state-size in a first pass, right before allocating the integrator SoA buffers in a second pass.
Reviewed By: brecht
Differential Revision: https://developer.blender.org/D16313
sycl::info::device::ext_intel_* descriptors are deprecated,
replaced with sycl::ext::intel::info::device:: that are available from
6.0+, for which we now check version in CMake.
JIT compilation of oneAPI kernels now happens during load stage
and proper message gets shown in the GUI during compilation.
Also, this implementation skips kernels that aren't needed for
the used scene, reducing overall (re)compilation time.
This is a minimal set of changes, allowing a lot of cleanup that can
happen afterward as it allows sycl method and objects to be used outside
of kernel.cpp.
Reviewed By: brecht, sergey
Differential Revision: https://developer.blender.org/D15397
When verbose level 4 is enabled, Blender prints kernel performance
data for Cycles on GPU backends (except Metal that doesn't use
debug_enqueue_* methods) for groups of kernels.
These changes introduce a new CYCLES_DEBUG_PER_KERNEL_PERFORMANCE
environment variable to allow getting timings for each kernels
separately and not grouped with others. This is done by adding
explicit synchronization after each kernel execution.
Differential Revision: https://developer.blender.org/D15971
Additionally, just stick to a pure error stating. Such messages
are aimed for developers and it is rather implied that oneAPI
rendering will be disabled.
The number of Execution Units and resident "threads" (simd width * threads
per EUs) are now exposed and used to select the number of states using
a simplified heuristic.
