Age | Commit message (Collapse) | Author |
|
In some instances, the number of control vertices of a hair could change mid-frame.
Cycles would then be unable to calculate proper motion blur for those hairs. This adds
interpolated CVs to fill in for the missing data. While this will not necessarily result in
a fully accurate reconstruction of the guide hair, it preserves motion blur instead of disabling it.
Reviewers: #cycles, sergey
Reviewed By: #cycles, sergey
Subscribers: sergey, brecht, #cycles
Tags: #cycles
Differential Revision: https://developer.blender.org/D3695
|
|
process
With small tiles, the repeated allocations on GPUs can actually slow down the denoising quite a lot.
Allocating the buffer just once reduces rendertime for the default cube with 16x16 tiles and denoising on a mobile 1050 from 22.7sec to 14.0sec.
|
|
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
|
|
|
|
This deduplicates the calls for tile (un)mapping and allows to have a target buffer that is different from the source buffer (needed for baking and animation denoising).
|
|
|
|
Work around what might be a compiler bug.
|
|
Reviewers: sergey, lukasstockner97
Reviewed By: lukasstockner97
Tags: #cycles, #bf_blender
Differential Revision: https://developer.blender.org/D3472
|
|
|
|
|
|
|
|
This should be the last Fermi removal commit, unless I missed something.
It's been a pleasure Fermi!
|
|
device_cuda.cpp.
Fermi code in Cycles kernel and texture system are coming next.
|
|
This brings separate initialization for libcuda and libnvrtc, which
fixes Cycles nvrtc compilation not working on build machines without
CUDA hardware available.
Differential Revision: https://developer.blender.org/D3045
|
|
nvcc is very picky regarding compiler versions, severely limiting the compiler we can use, this commit adds a nvrtc based compiler that'll allow us to build the cubins even if the host compiler is unsupported. for details see D2913.
Differential Revision: http://developer.blender.org/D2913
|
|
This was we can introduce other types of BVH, for example, wider ones, without
causing too much mess around boolean flags.
Thoughs:
- Ideally device info should probably return bitflag of what BVH types it
supports.
It is possible to implement based on simple logic in device/ and mesh.cpp,
rest of the changes will stay the same.
- Not happy with workarounds in util_debug and duplicated enum in kernel.
Maybe enbum should be stores in kernel, but then it's kind of weird to include
kernel types from utils. Soudns some cyclkic dependency.
Reviewers: brecht, maxim_d33
Reviewed By: brecht
Differential Revision: https://developer.blender.org/D3011
|
|
This is not expected to fix all issues. Also adds some more details
to error reporting to investigate failures.
|
|
In that case it can now fall back to CPU memory, at the cost of reduced
performance. For scenes that fit in GPU memory, this commit should not
cause any noticeable slowdowns.
We don't use all physical system RAM, since that can cause OS instability.
We leave at least half of system RAM or 4GB to other software, whichever
is smaller.
For image textures in host memory, performance was maybe 20-30% slower
in our tests (although this is highly hardware and scene dependent). Once
other type of data doesn't fit on the GPU, performance can be e.g. 10x
slower, and at that point it's probably better to just render on the CPU.
Differential Revision: https://developer.blender.org/D2056
|
|
Fixes a few corner cases found while stress testing host mapped memory.
|
|
|
|
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.
|
|
Reviewers: sergey, brecht
Differential Revision: https://developer.blender.org/D2920
|
|
|
|
|
|
Goal is to reduce OpenCL kernel recompilations.
Currently viewport renders are still set to use 64 closures as this seems to
be faster and we don't want to cause a performance regression there. Needs
to be investigated.
Reviewed By: brecht
Differential Revision: https://developer.blender.org/D2775
|
|
It should not be needed as far as I know, but just in case it fixes any
of the recent issues like T52572.
|
|
|
|
This way we can log the amount of memory used, and it will be important
for host mapped memory support.
|
|
This is a prequisite for getting host memory allocation to work. There appears
to be no support for 3D textures using host memory. The original version of
this code was written by Stefan Werner for D2056.
|
|
|
|
* Remove tex_* and pixels_* functions, replace by mem_*.
* Add MEM_TEXTURE and MEM_PIXELS as memory types recognized by devices.
* No longer create device_memory and call mem_* directly, always go
through device_only_memory, device_vector and device_pixels.
|
|
|
|
|
|
|
|
|
|
This change affects CUDA GPUs not connected to a display or connected to a
display but supporting compute preemption so that the display does not
freeze. I couldn't find an official list, but compute preemption seems to be
only supported with GTX 1070+ and Linux (not GTX 1060- or Windows).
This helps improve small tile rendering performance further if there are
sufficient samples x number of pixels in a single tile to keep the GPU busy.
|
|
|
|
* Use common TextureInfo struct for all devices, except CUDA fermi.
* Move image sampling code to kernels/*/kernel_*_image.h files.
* Use arrays for data textures on Fermi too, so device_vector<Struct> works.
|
|
|
|
|
|
OpenCL.
The work size is still very conservative, and this doesn't help for progressive
refine. For that we will need to render multiple tiles at the same time. But this
should already help for denoising renders that require too much memory with big
tiles, and just generally soften the performance dropoff with small tiles.
Differential Revision: https://developer.blender.org/D2856
|
|
There is no significant difference in denoised benchmark scenes and
denoising ctests, so might as well make it all consistent.
|
|
A little faster on some benchmark scenes, a little slower on others, seems
about performance neutral on average and saves a little memory.
|
|
This makes sharing some code between mega/split in following commits a bit
easier, and also paves the way for rendering multiple tiles later.
|
|
Makes it possible to call a function like mem_alloc() when the context is
already active. Also fixes some missing pops in case of errors.
|
|
I don't know if this will actually work, needs testing. Ref T52064.
|
|
Image textures were being packed into a single buffer for OpenCL, which
limited the amount of memory available for images to the size of one
buffer (usually 4gb on AMD hardware). By packing textures into multiple
buffers that limit is removed, while simultaneously reducing the number
of buffers that need to be passed to each kernel.
Benchmarks were within 2%.
Fixes T51554.
Differential Revision: https://developer.blender.org/D2745
|
|
On Pabellon, 25.8s mega, 35.4s split before, 32.7s split after.
|
|
Some of the functions might have been inlined, but others i don't see
how that was possible (don't think virtual functions can be inlined here).
In any case, better be explicitly optimal in the code.
|