Age | Commit message (Collapse) | Author |
|
|
|
This keeps render results compatible for combined CPU + GPU rendering.
Peformance and quality primitives is quite different than before. There
are now two options:
* Rounded Ribbon: render hair as flat ribbon with (fake) rounded normals, for
fast rendering. Hair curves are subdivided with a fixed number of user
specified subdivisions.
This gives relatively good results, especially when used with the Principled
Hair BSDF and hair viewed from a typical distance. There are artifacts when
viewed closed up, though this was also the case with all previous primitives
(but different ones).
* 3D Curve: render hair as 3D curve, for accurate results when viewing hair
close up. This automatically subdivides the curve until it is smooth.
This gives higher quality than any of the previous primitives, but does come
at a performance cost and is somewhat slower than our previous Thick curves.
The main problem here is performance. For CPU and OpenCL rendering performance
seems usually quite close or better for similar quality results.
However for CUDA and Optix, performance of 3D curve intersection is problematic,
with e.g. 1.45x longer render time in Koro (though there is no equivalent quality
and rounded ribbons seem fine for that scene). Any help or ideas to optimize this
are welcome.
Ref T73778
Depends on D8012
Maniphest Tasks: T73778
Differential Revision: https://developer.blender.org/D8013
|
|
Triangles were very memory intensive. The only reason they were not removed yet
is that they gave more accurate results, but there will be an accurate 3D curve
primitive added for this.
Line rendering was always poor quality since the ends do not match up. To keep CPU
and GPU compatibility we just remove them entirely. They could be brought back if
an Embree compatible implementation is added, but it's not clear to me that there
is a use case for these that we'd consider important.
Ref T73778
Reviewers: #cycles
Subscribers:
|
|
|
|
bright objects
The input data to the OptiX denoiser was clamped to 0..10000 as required, but it could easily
exceed that range with a high number of samples (since the data contains the overall sum). To
fix that, divide by the number of samples first and multiply it back in after the denoiser ran.
|
|
There should be no user visible change from this, except that tile size
now affects performance. The goal here is to simplify bake denoising in
D3099, letting it reuse more denoising tiles and pass code.
A lot of code is now shared with regular rendering, with the two main
differences being that we read some render result passes from the bake API
when starting to render a tile, and call the bake kernel instead of the
path trace kernel.
With this kind of design where Cycles asks for tiles from the bake API,
it should eventually be easier to reduce memory usage, show tiles as
they are baked, or bake multiple passes at once, though there's still
quite some work needed for that.
Reviewers: #cycles
Subscribers: monio, wmatyjewicz, lukasstockner97, michaelknubben
Differential Revision: https://developer.blender.org/D3108
|
|
This is not yet fully supported by automatic volume bounds but works fine in
most cases that will have mostly matching bounds.
Ref T73201
|
|
There was too much image texture specific stuff in device_memory, and too
much code duplication between devices.
|
|
This is legacy code from when we had a fixed number of textures.
|
|
|
|
This feature takes some inspiration from
"RenderMan: An Advanced Path Tracing Architecture for Movie Rendering" and
"A Hierarchical Automatic Stopping Condition for Monte Carlo Global Illumination"
The basic principle is as follows:
While samples are being added to a pixel, the adaptive sampler writes half
of the samples to a separate buffer. This gives it two separate estimates
of the same pixel, and by comparing their difference it estimates convergence.
Once convergence drops below a given threshold, the pixel is considered done.
When a pixel has not converged yet and needs more samples than the minimum,
its immediate neighbors are also set to take more samples. This is done in order
to more reliably detect sharp features such as caustics. A 3x3 box filter that
is run periodically over the tile buffer is used for that purpose.
After a tile has finished rendering, the values of all passes are scaled as if
they were rendered with the full number of samples. This way, any code operating
on these buffers, for example the denoiser, does not need to be changed for
per-pixel sample counts.
Reviewed By: brecht, #cycles
Differential Revision: https://developer.blender.org/D4686
|
|
The OptiX denoiser can be a great help when rendering in the viewport, since it is really fast
and needs few samples to produce convincing results. This patch therefore adds support for
using any Cycles denoiser in the viewport also (but only the OptiX one is selectable because
the NLM one is too slow to be usable currently). It also adds support for denoising on a
different device than rendering (so one can e.g. render with the CPU but denoise with OptiX).
Reviewed By: #cycles, brecht
Differential Revision: https://developer.blender.org/D6554
|
|
This patch adds support for the OptiX denoiser as an alternative to the existing NLM denoiser in Cycles. It's re-using the same denoising architecture based on tiles and therefore implicitly also works with multiple GPUs.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D6395
|
|
The OptiX intersection program for curves uses "optixGetObjectRayDirection"
to get the ray direction in object space (which was inverse transformed
with the current transformation matrix). OptiX does no additional operations
on it, so if there is a scaling transform, the direction is not normalized.
But the curve intersection routine expects that. In addition, the distances
used in "optixGetRayTmax()" and "optixReportIntersection()" are in world
space, so need to adjust them accordingly.
|
|
This adds all the kernel side changes for the Optix backend.
Ref D5363
|
|
Differential Revision: https://developer.blender.org/D5326
|
|
|
|
Apply clang format as proposed in T53211.
For details on usage and instructions for migrating branches
without conflicts, see:
https://wiki.blender.org/wiki/Tools/ClangFormat
|
|
Part of the cleanup of the OpenCL codebase.
Single program is not effective when using OpenCL, it is slower
to compile and slower during rendering (when used in for example
`barbershop` or `victor`).
Reviewers: brecht, #cycles
Maniphest Tasks: T62267
Differential Revision: https://developer.blender.org/D4481
|
|
The goal of this patch is to have limit the number of times
kernels needs to be compiled and are reused as kernels with
different compile directives can lead to identical same
binaries.
The implementation does this by stripping the compile directives.
and reshuffling kernels so the output is more likely to be the
same.
We focussed on the kernels where it was easy to detect and maintain
(bundle, bake, displace, do_volume and background). More optimizations
could be done but they are probably less obvious.
Merged the data_init and state_buffer_size kernels to split_bundle.
This patch will also remove empty kernels for do_volume and bake
when their features are not enabled.
When using the benchmark files there are less background, bake and
do_volume kernels compiled.
Fix: T61576, T61501, T61466
Reviewed By: brecht, #cycles
Differential Revision: https://developer.blender.org/D4390
|
|
|
|
The bake kernels are also used during mesh displacement and light
importance sampling. We disabled the implementation of these kernels
when baking was not enabled.
|
|
Using OpenCL MegaKernel has been slow and therefore not usefull.
This patch will remove the mega kernel from the OpenCL codebase
and the OpenCLDeviceBase class.
T61736: removal of mega kernel
T61703: baking does not work with mega kernel
Tags: #cycles
Differential Revision: https://developer.blender.org/D4383
|
|
Cycles OpenCL: Split baking kernels in own program
Fix T61463. Before this patch baking was part of the base kernels. There
are 3 baking kernels that and all 3 uses shader evaluation. Only for one
of these kernels the functionality was wrapped in the __NO_BAKING__
compile directive.
When you start baking this leads to long compile times. By separating
in individual programs will reduce the compile times.
Also wrapped all baking kernels with __NO_BAKING__ to reduce the
compilation times.
Impact on compilation time
job | scene_name | previous | new | percentage
--------+-----------------+----------+-------+------------
T61463 | empty | 10.63 | 7.27 | 32%
T61463 | bmw | 17.91 | 14.24 | 20%
T61463 | fishycat | 19.57 | 15.08 | 23%
T61463 | barbershop | 54.10 | 48.18 | 11%
T61463 | classroom | 17.55 | 14.42 | 18%
T61463 | koro | 18.92 | 17.15 | 9%
T61463 | pavillion | 17.43 | 14.23 | 18%
T61463 | splash279 | 16.48 | 15.33 | 7%
T61463 | volume_emission | 36.22 | 34.19 | 6%
Impact on render time
job | scene_name | previous | new | percentage
--------+-----------------+----------+---------+------------
T61463 | empty | 21.06 | 20.54 | 2%
T61463 | bmw | 198.44 | 189.59 | 4%
T61463 | fishycat | 394.20 | 388.50 | 1%
T61463 | barbershop | 1188.16 | 1185.49 | 0%
T61463 | classroom | 341.08 | 339.27 | 1%
T61463 | koro | 472.43 | 360.70 | 24%
T61463 | pavillion | 905.77 | 902.14 | 0%
T61463 | splash279 | 55.26 | 54.92 | 1%
T61463 | volume_emission | 62.59 | 39.09 | 38%
I don't have a grounded explanation why koro and volume_emission is this much
faster; I have done several tests though...
Maniphest Tasks: T61463
Differential Revision: https://developer.blender.org/D4376
|
|
This patch implements a workaround to get the multithreaded compilation from D2231 working.
So far, it only works for Blender, not for Cycles Standalone. Also, I have only tested the Linux codepath in the helper function.
Depends on D2231.
Patch by lukasstockner97, jbakker, brecht
job | scene_name | compilation_time
----------+-----------------+------------------
Baseline | empty | 22.73
D2264 | empty | 13.94
Baseline | bmw | 56.44
D2264 | bmw | 41.32
Baseline | fishycat | 59.50
D2264 | fishycat | 45.19
Baseline | barbershop | 212.28
D2264 | barbershop | 169.81
Baseline | victor | 67.51
D2264 | victor | 53.60
Baseline | classroom | 51.46
D2264 | classroom | 39.02
Baseline | koro | 62.48
D2264 | koro | 49.03
Baseline | pavillion | 54.37
D2264 | pavillion | 38.82
Baseline | splash279 | 47.43
D2264 | splash279 | 37.94
Baseline | volume_emission | 145.22
D2264 | volume_emission | 121.10
This patch reduced compilation time as the split kernels and base
kernels are compiled in parallel. In cycles debug mode (256) you can set
unmark the opencl single program file, what reduces the compilation time
even further (bmw 17 seconds, barbershop 53 seconds).
Reviewers: brecht, dingto, sergey, juicyfruit, lukasstockner97
Reviewed By: brecht
Subscribers: Loner, jbakker, candreacchio, 3dLuver, LazyDodo, bliblubli
Differential Revision: https://developer.blender.org/D2264
|
|
This is the internal implementation, not available from the API or
interface yet. The algorithm takes into account past and future frames,
both to get more coherent animation and reduce noise.
Ref D3889.
|
|
Prefiltering of feature passes will happen during rendering, which can
then be used for denoising immediately or written as a render pass for
later (animation) denoising.
The number of denoising data passes written is reduced because of this,
leaving out the feature variance passes. The passes are now Normal,
Albedo, Depth, Shadowing, Variance and Intensity.
Ref D3889.
|
|
There may still be rendering errors when used for older graphics cards.
|
|
It was used in like 95% of places.
|
|
It is supposed to be two spaces before comment stating which if
else/endif statements corresponds to. Was mainly violated in the
header guards.
|
|
This allows for extra output passes that encode automatic object and material masks
for the entire scene. It is an implementation of the Cryptomatte standard as
introduced by Psyop. A good future extension would be to add a manifest to the
export and to do plenty of testing to ensure that it is fully compatible with other
renderers and compositing programs that use Cryptomatte.
Internally, it adds the ability for Cycles to have several passes of the same type
that are distinguished by their name.
Differential Revision: https://developer.blender.org/D3538
|
|
denoiser
Previously the code allocated its own temporary memory, but it's possible to just use the existing shared one instead.
|
|
|
|
cl_khr_fp16 extension.
|
|
|
|
Textures in 16 bit integer format are sometimes used for displacement, bump and normal maps and can be exported by tools like Substance Painter. Without this patch, Cycles would promote those textures to single precision floating point, causing them to take up twice as much memory as needed.
Reviewers: #cycles, brecht, sergey
Reviewed By: #cycles, brecht, sergey
Subscribers: sergey, dingto, #cycles
Tags: #cycles
Differential Revision: https://developer.blender.org/D3523
|
|
standard and get rid of set_tile_info
|
|
|
|
assuming sRGB
I've limited it to just the RGB<->XYZ stuff for now, correct image handling is the next step.
Reviewers: brecht, sergey
Differential Revision: https://developer.blender.org/D3478
|
|
This commit contains the minimum to make clang build/work with blender, asan and ninja build support is forthcoming
Things to note:
1) Builds and runs, and is able to pass all tests (except for the freestyle_stroke_material.blend test which was broken at that time for all platforms by the looks of it)
2) It's slightly faster than msvc when using cycles. (time in seconds, on an i7-3370)
victor_cpu
msvc:3099.51
clang:2796.43
pavillon_barcelona_cpu
msvc:1872.05
clang:1827.72
koro_cpu
msvc:1097.58
clang:1006.51
fishy_cat_cpu
msvc:815.37
clang:722.2
classroom_cpu
msvc:1705.39
clang:1575.43
bmw27_cpu
msvc:552.38
clang:561.53
barbershop_interior_cpu
msvc:2134.93
clang:1922.33
3) clang on windows uses a drop in replacement for the Microsoft cl.exe (takes some of the Microsoft parameters, but not all, and takes some of the clang parameters but not all) and uses ms headers + libraries + linker, so you still need visual studio installed and will use our existing vc14 svn libs.
4) X64 only currently, X86 builds but crashes on startup.
5) Tested with llvm/clang 6.0.0
6) Requires visual studio integration, available at https://github.com/LazyDodo/llvm-vs2017-integration
7) The Microsoft compiler spawns a few copies of cl in parallel to get faster build times, clang doesn't, so the build time is 3-4x slower than with msvc.
8) No openmp support yet. Have not looked at this much, the binary distribution of clang doesn't seem to include it on windows.
9) No ASAN support yet, some of the sanitizers can be made to work, but it was decided to leave support out of this commit.
Reviewers: campbellbarton
Differential Revision: https://developer.blender.org/D3304
|
|
|
|
This should be the last Fermi removal commit, unless I missed something.
It's been a pleasure Fermi!
|
|
Did not touch Texture related defines, that comes next.
|
|
Not sure why it was in there, all the debug flags stuff is to be handled outside
of kernel.
|
|
Our own implementation was behaving different comparing to OSL and GPU,
namely on the border pixels OSL and CUDA was doing interpolation with
black, but we were clamping coordinate.
This partially fixes issue reported in T53452.
Similar change should also be done for 3D interpolation perhaps, but this
is to be investigated separately.
|
|
|
|
Previously, the NLM kernels would be launched once per offset with one thread per pixel.
However, with the smaller tile sizes that are now feasible, there wasn't enough work to fully occupy GPUs which results in a significant slowdown.
Therefore, the kernels are now launched in a single call that handles all offsets at once.
This has two downsides: Memory accesses to accumulating buffers are now atomic, and more importantly, the temporary memory now has to be allocated for every shift at once, increasing the required memory.
On the other hand, of course, the smaller tiles significantly reduce the size of the memory.
The main bottleneck right now is the construction of the transformation - there is nothing to be parallelized there, one thread per pixel is the maximum.
I tried to parallelize the SVD implementation by storing the matrix in shared memory and launching one block per pixel, but that wasn't really going anywhere.
To make the new code somewhat readable, the handling of rectangular regions was cleaned up a bit and commented, it should be easier to understand what's going on now.
Also, some variables have been renamed to make the difference between buffer width and stride more apparent, in addition to some general style cleanup.
|
|
|
|
CUDA 9.0.176 apparently caused some slow down on high-end Pascal cards that can be mitigated by increasing the number of registers. See https://developer.blender.org/F1142667 for a detailed comparison.
|
|
|