It will discard the whole tile, but it's still kind of more friendly than
fully locked interface (sort of) for until tile is fully sampled.
Sorry if it causes PITA to merge for the opencl split work, but this issue
bothering a lot when collecting benchmarks.
Basically just moves cached kernels from ~/.config/blender/BLENDER_VERSION to
~/.cache/cycles/kernels. This has following benefits:
- Follows XDG specification more closely,
not as if it's totally crucial or measurable by users, but still nice.
- Prevents unexpected sizes of config folder, makes disk space used in more
predictable for users way.
- Allows to share kernels across multiple Blender versions,
which makes it easier debugging at the times close to release.
- "Copy Previous Settings" operator will no longer be copying possibly
gigabytes of cached kernels, which used to lead to really nast disk usage
and annoying delays of copying settings.
- In the future we can have some smart logic to clear old unused cached
kernels.
Currently only done for Linux and OSX. Windows still follows old "cache"
folder logic, but it's not really important for now because we don't
support kernel compilation on this platform yet.
Reviewers: dingto, juicyfruit, brecht
Reviewed By: brecht
Differential Revision: https://developer.blender.org/D2197
Kernels can now be built without patch evaluation when not needed by the
scene (Catmull-Clark subdivision not in use), giving a performance boost
for some devices.
It is possible that compilation will fail without giving anything in the
log buffer. For this cases giving a tip about error code will be really
handy.
Patch by @Ilia, thanks!
This way we can easily switch between toolkits without worrying
whether some kernel was compiled with old or new CUDA toolkit.
It's also now possible to switch machine architecture and have
proper cached kernel detected. Not as if it happens every day,
but i did such a bitness switch back in the days :)
All the changes are mainly giving explicit tips on inlining functions,
so they match how inlining worked with previous toolkit.
This make kernel compiled by CUDA 8 render in average with same speed
as previous kernels. Some scenes are somewhat faster, some of them are
somewhat slower. But slowdown is within 1% so far.
On a positive side it allows us to enable newer generation cards on
buildbots (so GTX 10x0 will be officially supported soon).
Some of these values can get quite large and are hard to read, adding this
makes it easy to read them at a glance.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D2039
This adds support for CUDA Texture objects (also known as Bindless textures) for Kepler GPUs (Geforce 6xx and above).
This is used for all 2D/3D textures, data still uses arrays as before.
User benefits:
* No more limits of image textures on Kepler.
We had 5 float4 and 145 byte4 slots there before, now we have 1024 float4 and 1024 byte4.
This can be extended further if we need to (just change the define).
* Single channel textures slots (byte and float) are now supported on Kepler as well (1024 slots for each type).
ToDo / Issues:
* 3D textures don't work yet, at least don't show up during render. I have no idea whats wrong yet.
* Dynamically allocate bindless_mapping array?
I hope Fermi still works fine, but that should be tested on a Fermi card before pushing to master.
Part of my GSoC 2016.
Reviewers: sergey, #cycles, brecht
Subscribers: swerner, jtheninja, brecht, sergey
Differential Revision: https://developer.blender.org/D1999
This commit makes it so malloc() is only happening once per volume and
once per transparent shadow query (per thread), improving scalability of
the code to multiple CPU cores.
Hard to measure this with a low-bottom i7 here currently, but from quick
tests seems volume sampling gave about 3-5% speedup.
The idea is to store allocated memory in kernel globals, which are per
thread on CPU already.
Reviewers: dingto, juicyfruit, lukasstockner97, maiself, brecht
Reviewed By: brecht
Subscribers: Blendify, nutel
Differential Revision: https://developer.blender.org/D1996
Until now, single channel textures were packed into a float4, wasting 3 floats per pixel. Memory usage of such textures is now reduced by 3/4.
Voxel Attributes such as density, flame and heat benefit from this, but also Bumpmaps with one channel.
This commit also includes some cleanup and code deduplication for image loading.
Example Smoke render from Cosmos Laundromat: http://www.pasteall.org/pic/show.php?id=102972
Memory here went down from ~600MB to ~300MB.
Reviewers: #cycles, brecht
Differential Revision: https://developer.blender.org/D1981
* When Baking wasn't used we got an error.
* On top of Volume Nodes (NODES_FEATURE_VOLUME), we now also check if we need volume sampling code,
so we can disable that as well and save some further compilation time.
If the CUDA Toolkit is installed and the user is on Linux,
adaptive, feature based CUDA runtime compile is now possible to enable via:
* Environment flag CYCLES_CUDA_ADAPTIVE_COMPILE or
* Debug menu (Debug value 256) in the Cycles UI.
This was a hard decision, because going newer CUDA toolkit makes
rendering up to 5% slower. But on another hand, it solves major
speed regressions (up to 30%) with branched path tracing on a
top level cards.
Neither of those regressions have a meaningful and sane workaround
from the code itself.
Toolkit 6.5 could still be used, but it's no longer recommended one.
Supports both smoke/fire and point density textures now.
Reduces number of textures available for sm_20 and sm_21, but you have
to compromise somewhere on such a limited hardware.
Currently limited to linear interpolation only, and decoupled ray
marching is not supported yet. Think those could be considered just a
further improvement.
Some quick example:
https://developer.blender.org/F282934
Code is minimal and we can fully consider it a fix for missing
support of 3D textures with CUDA.
Reviewers: lukasstockner97, brecht, juicyfruit, dingto
Reviewed By: brecht, juicyfruit, dingto
Subscribers: mib2berlin
Differential Revision: https://developer.blender.org/D1806
We don't have vectors re-allocation happening multiple times from inside
a loop anymore, so we can safely switch to a memory guarded allocator for
vectors and keep track on the memory usage at various stages of rendering.
Additionally, when building from inside Blender repository, Cycles will
use Blender's guarded allocator, so actual memory usage will be displayed
in the Space Info header.
There are couple of tricky aspects of the patch:
- TaskScheduler::exit() now explicitly frees memory used by `threads`.
This is needed because `threads` is a static member which destructor
isn't getting called on Blender's exit which caused memory leak print
to happen.
This shouldn't give any measurable speed issues, reallocation of that
vector is only one of fewzillion other allocations happening during
synchronization.
- Use regular guarded malloc (not aligned one). No idea why it was
made to be aligned in the first place. Perhaps some corner case tests
or so. Vector was never expected to be aligned anyway. Let's see if
we'll have actual bugs with this.
Reviewers: dingto, lukasstockner97, juicyfruit, brecht
Reviewed By: brecht
Differential Revision: https://developer.blender.org/D1774
Basically the idea is to make code robust against extending
enum options in the future by falling back to a known safe
default setting when RNA is set to something unknown.
While this approach solves the issues similar to T47377,
but it wouldn't really help when/if any of the RNA values
gets ever deprecated and removed. There'll be no simple
solution to that apart from defining explicit mapping from
RNA value to Cycles one.
Another part which isn't so great actually is that we now
have to have some enum guards and give some explicit values
to the enum items, but we can live with that perhaps.
Reviewers: dingto, juicyfruit, lukasstockner97, brecht
Reviewed By: brecht
Differential Revision: https://developer.blender.org/D1785
The title says it all actually, the idea is to make Cycles
only requiring Boost via 3rd party dependencies like OIIO
and OSL.
So now there are only few places which still uses Boost:
- Foreach, function bindings and threading primitives.
Those we can easily get rid with C++11 bump (which seems
inevitable sooner or later if we'll want ot use newer
LLVM for OSL),
- Networking devices
There's no quick solution for those currently, but there
are some patches around which improves serialization.
Reviewers: juicyfruit, mont29, campbellbarton, brecht, dingto
Reviewed By: brecht, dingto
Differential Revision: https://developer.blender.org/D1764
The idea is to switch from allocating separate buffers for shader data's
structure of arrays to allocating one huge memory block and do some index
trickery to make it accessed as SOA.
This saves quite reasonable amount of lines of code in device_opencl and
also makes it possible to get rid of special declaration of ShaderData
structure.
As a side effect it also makes it easier to experiment with SOA vs. AOS
for split kernel.
Works fine here on NVidia GTX580, Intel CPU amd AMD Fiji cards.
Reviewers: #cycles, brecht, juicyfruit, dingto
Differential Revision: https://developer.blender.org/D1593
Use KernelGlobals to access all the global arrays for the intermediate
storage instead of passing all this storage things explicitly.
Tested here with Intel OpenCL, NVIDIA GTX580 and AMD Fiji, didn't see
any artifacts, so guess it's all good.
Reviewers: juicyfruit, dingto, lukasstockner97
Differential Revision: https://developer.blender.org/D1736
There is no function pointers in OpenCL specification. For as long
as we want to support this platform we should follow the specifications.
While the code is not totally optimal now, it should not be that huge
of performance issue on CPU since it does jump tables just nicely, so
it's not that much extra computation here.
