* Rename struct KernelGlobals to struct KernelGlobalsCPU
* Add KernelGlobals, IntegratorState and ConstIntegratorState typedefs
that every device can define in its own way.
* Remove INTEGRATOR_STATE_ARGS and INTEGRATOR_STATE_PASS macros and
replace with these new typedefs.
* Add explicit state argument to INTEGRATOR_STATE and similar macros
In preparation for decoupling main and shadow paths.
Differential Revision: https://developer.blender.org/D12888
This is the first of a sequence of changes to support compiling Cycles kernels as MSL (Metal Shading Language) in preparation for a Metal GPU device implementation.
MSL requires that all pointer types be declared with explicit address space attributes (device, thread, etc...). There is already precedent for this with Cycles' address space macros (ccl_global, ccl_private, etc...), therefore the first step of MSL-enablement is to apply these consistently. Line-for-line this represents the largest change required to enable MSL. Applying this change first will simplify future patches as well as offering the emergent benefit of enhanced descriptiveness.
The vast majority of deltas in this patch fall into one of two cases:
- Ensuring ccl_private is specified for thread-local pointer types
- Ensuring ccl_global is specified for device-wide pointer types
Additionally, the ccl_addr_space qualifier can be removed. Prior to Cycles X, ccl_addr_space was used as a context-dependent address space qualifier, but now it is either redundant (e.g. in struct typedefs), or can be replaced by ccl_global in the case of pointer types. Associated function variants (e.g. lcg_step_float_addrspace) are also redundant.
In cases where address space qualifiers are chained with "const", this patch places the address space qualifier first. The rationale for this is that the choice of address space is likely to have the greater impact on runtime performance and overall architecture.
The final part of this patch is the addition of a metal/compat.h header. This is partially complete and will be extended in future patches, paving the way for the full Metal implementation.
Ref T92212
Reviewed By: brecht
Maniphest Tasks: T92212
Differential Revision: https://developer.blender.org/D12864
The assumption about absent shadow path was wrong.
The rest of the changes are to ensure shadow paths are finished prior
to the split, so that they write to the proper passes.
The issue was caught by running regression tests on OptiX.
Differential Revision: https://developer.blender.org/D12857
For details see the "Extending the Disney BRDF to a BSDF with Integrated
Subsurface Scattering" paper.
We split the diffuse BSDF into a lambertian and retro-reflection component.
The retro-reflection component is always handled as a BSDF, while the
lambertian component can be replaced by a BSSRDF.
For the BSSRDF case, we compute Fresnel separately at the entry and exit
points, which may have different normals. As the scattering radius decreases
this converges to the BSDF case.
A downside is that this increases noise for subsurface scattering in the
Principled BSDF, due to some samples going to the retro-reflection component.
However the previous logic (also in 2.93) was simple wrong, using a
non-sensical view direction vector at the exit point. We use an importance
sampling weight estimate for the retro-reflection to try to better balance
samples between the BSDF and BSSRDF.
Differential Revision: https://developer.blender.org/D12801
There is not enough time before the release to improve Random Walk to handle
all cases this was used for, so restore it for now.
Since there is no more path splitting in cycles-x, this can increase noise in
non-flat areas for the sample number of samples, though fewer rays will be traced
also. This is fundamentally a trade-off we made in the new design and why Random
Walk is a better fit. However the importance resampling we do now does help to
reduce noise.
Differential Revision: https://developer.blender.org/D12800
Only copy required part of volume stack instead of entire stack.
Solves time regression introduced by D12759 and avoids need in
implementing volume stack calculation to exactly match what the
path tracing will do (as well as potentially makes scenes with
a lot of volumes ans a tiny bit of deeply nested ones render
faster).
Still need to look into memory aspect of the regression, but
that is for separate patch.
Ref T92014
Maniphest Tasks: T92014
Differential Revision: https://developer.blender.org/D12790
Previously the storage here was optimized to avoid indirections in BVH2
traversal. This helps improve performance a bit, but makes performance
and memory usage of Embree and OptiX BVHs a bit worse also. It also adds
code complexity in other parts of the code.
Now decouple triangle and curve primitive storage from BVH2.
* Reduced peak memory usage on all devices
* Bit better performance for OptiX and Embree
* Bit worse performance for CUDA
* Simplified code:
** Intersection.prim/object now matches ShaderData.prim/object
** No more offset manipulation for mesh displacement before a BVH is built
** Remove primitive packing code and flags for Embree and OptiX
** Curve segments are now stored in a KernelCurve struct
* Also happens to fix a bug in baking with incorrect prim/object
Fixes T91968, T91770, T91902
Differential Revision: https://developer.blender.org/D12766
MSVC does not support variable size array definition.
Use maximum possible stack, similar to the GPU case.
Not expected to have user-measurable difference.
Make volume stack allocated conditionally, potentially based on the
actual nested level of objects in the scene.
Currently the nested level is estimated by number of volume objects.
This is a non-expensive check which is probably enough in practice
to get almost perfect memory usage and performance.
The conditional allocation is a bit tricky.
For the CPU we declare and define maximum possible volume stack,
because there are only that many integrator states on the CPU.
On the GPU we declare outer SoA to have all volume stack elements,
but only allocate actually needed ones. The actually used volume
stack size is passed as a pre-processor, which seems to be easiest
and fastest for the GPU state copy.
There seems to be no speed regression in the demo files on RTX6000.
Note that scenes with high nested level of volume will now be slower
but correct.
Differential Revision: https://developer.blender.org/D12759
Fixes:{T91064}
Caused by {rBcd118c5581f482afc8554ff88b5b6f3b552b1682}
- Applies `ensure_valid_reflection()` to the normal input on all BSDFs for CPU and GPU.
- This doesn't affect hair.
- Removes `ensure_valid_reflection()` from the output of Bump Map and Normal Map nodes for CPU/GPU as it is not needed.
- The fix doesn't touch OSL.
Reviewed By: brecht, leesonw
Maniphest Tasks: T91064
Differential Revision: https://developer.blender.org/D12403
This effectively undoes some of the following commit:
rB4537e8558468c71a03bf53f59c60f888b3412de2
The tables in question were duplicated 5-6 times into the blender
executable due to the headers being used in multiple translation units.
This contributes ~6.3kb worth of duplicate data into the binary.
Some further details are in the below revision.
Differential Revision: https://developer.blender.org/D12724
Implement an overscan support for tiles, so that adaptive sampling can
rely on the pixels neighbourhood.
Differential Revision: https://developer.blender.org/D12599
Always sample background pass behind shadow catcher (if the pass
exists, of course), regardless of whether shadow catcher will be
used as approximate or accurate.
Allows to combine accurate shadows into an environment map.
Differential Revision: https://developer.blender.org/D12747
Similar to the previous change in the area: need to avoid ray
point and direction becoming a non-finite value.
Use the view direction when the geometrical normal can not be
calculated.
Collaboration and sanity inspiration with Brecht!
Differential Revision: https://developer.blender.org/D12703
The issue was caused by hair shader setup setting normal to a non
finite value, which then gets used to create a ray with non-finite
direction, making BVH traversal to run out of stack memory.
Happens with 150_0040_A.lighting.blend frame 112 of the Sprites
project.
Differential Revision: https://developer.blender.org/D12692
Avoids possible numerical issues in the path tracing kernel, which
is most important for displacement as non-finite values in BVH can
lead to infinite node recursion during traversal.
Differential Revision: https://developer.blender.org/D12690
NOTE: this feature is not ready for user testing, and not yet enabled in daily
builds. It is being merged now for easier collaboration on development.
HIP is a heterogenous compute interface allowing C++ code to be executed on
GPUs similar to CUDA. It is intended to bring back AMD GPU rendering support
on Windows and Linux.
https://github.com/ROCm-Developer-Tools/HIP.
As of the time of writing, it should compile and run on Linux with existing
HIP compilers and driver runtimes. Publicly available compilers and drivers
for Windows will come later.
See task T91571 for more details on the current status and work remaining
to be done.
Credits:
Sayak Biswas (AMD)
Arya Rafii (AMD)
Brian Savery (AMD)
Differential Revision: https://developer.blender.org/D12578
Before the visibility test against the visibility flags was performed in an any-hit program in OptiX
(called `__anyhit__kernel_optix_visibility_test`), which was using the `__prim_visibility` array.
This is not entirely correct however, since `__prim_visibility` is filled with the merged visibility
flags of all objects that reference that primitive, so if one object uses different visibility flags
than another object, but they both are instances of the same geometry, they would appear the same
way. The reason that the any-hit program was used rather than the OptiX instance visibility mask is
that the latter is currently limited to 8 bits only, which is not sufficient to contain all Cycles
visibility flags (12 bits).
To mostly fix the problem with multiple instances and different visibility flags, I changed things to
use the OptiX instance visibility mask for a subset of the Cycles visibility flags (`PATH_RAY_CAMERA`
to `PATH_RAY_VOLUME_SCATTER`, which fit into 8 bits) and only fall back to the visibility test any-hit
program if that isn't enough (e.g. the ray visibility mask exceeds 8 bits or when using the built-in
curves from OptiX, since the any-hit program is then also used to skip the curve endcaps).
This may also improve performance in some cases, since by default OptiX can now perform the normal
scene intersection trace calls entirely on RT cores without having to jump back to the SM on every
hit to execute the any-hit program.
Fixes T89801
Differential Revision: https://developer.blender.org/D12604
Fix T91632: Stops the sample correlation between dimensions which was causing rendering artefacts on simple scenes.
This is done by increasing the amount of jitter the Cranley Patterson Rotation is allowed to add. Also, it uses the y dimension of the of the sample table for 1D sampling which causes further decorrelation between dimensions. As an additional measure the x and y dimensions are swapped randomly to provide further decorrelation.
Maniphest Tasks: T91632
Differential Revision: https://developer.blender.org/D12610
Goal is to add the length attribute to the Hair Info node, for better control over color gradients or similar along the hair.
Reviewed By: #eevee_viewport, brecht
Differential Revision: https://developer.blender.org/D10481
