git.blender.org/blender.git - Unnamed repository; edit this file 'description' to name the repository.

Age	Commit message (Collapse)	Author
2017-03-08	Cycles: Make it more obvious message which initialization failed	Sergey Sharybin

2017-03-08	Cycles: Use 1-based line number for #line directives	Sergey Sharybin
	AMD CPU platform was complaining about #line 0 directives in the code.
2017-03-08	Cycles: Log which device kernels are being loaded for	Sergey Sharybin

2017-03-08	Cycles: Make it possible to access KernelGlobals from split data ↵	Sergey Sharybin
	initialization function
2017-03-08	Cycles: Cleanup, remove residue of previous split kernel data	Sergey Sharybin
	This is all in split data state array.
2017-03-08	Cycles: Fix indentation	Mai Lavelle

2017-03-08	Cycles: Fix strict warning about unused variable	Mai Lavelle

2017-03-08	Cycles: Calculate size of split state buffer kernel side	Mai Lavelle
	By calculating the size of the state buffer in the kernel rather than the host less code is needed and the size actually reflects the requested features. Will also be a little faster in some cases because of larger global work size.
2017-03-08	Cycles: Fix crash after failed kernel build	Mai Lavelle
	Pointers to kernels were uninitialized leading to freeing of random memory addresses. Another reason it would be good to use smart pointers.
2017-03-08	Cycles: Faster building of split kernel	Mai Lavelle
	Simple change to make it so that only kernels that have been modified are rebuilt. Might only be useful during development.
2017-03-08	Cycles: Initialize rng_state for split kernel	Mai Lavelle
	Because the split kernel can render multiple samples in parallel it is necessary to have everything initialized before rendering of any samples begins. The code that normally handles initialization of `rng_state` (`kernel_path_trace_setup()`) only does so for the first sample, which was causing artifacts in the split kernel due to uninitialized `rng_state` for some samples. Note that because the split kernel can render samples in parallel this means that the split kernel is incompatible with the LCG.
2017-03-08	Cycles: Remove sum_all_radiance kernel	Mai Lavelle
	This was only needed for the previous implementation of parallel samples. As we don't have that any more it can be removed. Real reason for removal tho is this: `per_sample_output_buffers` was being calculated too small and artifacts resulted. The tile buffer is already the correct size and calculating the size for `per_sample_output_buffers` is a bit difficult with the current layout of the code. As `per_sample_output_buffers` was only needed for `sum_all_radiance`, removing that kernel and writing output to the tile buffer directly fixes the artifacts.
2017-03-08	Cycles: Split path initialization into own kernel	Mai Lavelle
	This makes it easier to initialize things correctly in the data_init kernel before they are needed by path tracing.
2017-03-08	Cycles: Seperate kernel loading time from render time	Mai Lavelle

2017-03-08	Cycles: Add names to buffer allocations	Mai Lavelle
	This is to help debug and track memory usage for generic buffers. We have similar for textures already since those require a name, but for buffers the name is only for debugging proposes.
2017-03-08	Cycles: CUDA implementation of split kernel	Mai Lavelle

2017-03-08	Cycles: CPU implementation of split kernel	Mai Lavelle

2017-03-08	Cycles: Remove ccl_fetch and SOA	Mai Lavelle

2017-03-08	Cycles: Report device maximum allocation and detected global size	Sergey Sharybin

2017-03-08	Cycles: Workaround for driver hangs	Mai Lavelle
	Simple workaround for some issues we've been having with AMD drivers hanging and rendering systems unresponsive. Unfortunately this makes things a bit slower, but its better than having to do hard reboots. Will be removed when drivers have been fixed. Define CYCLES_DISABLE_DRIVER_WORKAROUNDS to disable for testing purposes.
2017-03-08	Cycles: OpenCL split kernel refactor	Mai Lavelle
	This does a few things at once: - Refactors host side split kernel logic into a new device agnostic class `DeviceSplitKernel`. - Removes tile splitting, a new work pool implementation takes its place and allows as many threads as will fit in memory regardless of tile size, which can give performance gains. - Refactors split state buffers into one buffer, as well as reduces the number of arguments passed to kernels. Means there's less code to deal with overall. - Moves kernel logic out of OpenCL kernel files so they can later be used by other device types. - Replaced OpenCL specific APIs with new generic versions - Tiles can now be seen updating during rendering
2017-03-08	Cycles: Add OpenCL kernel for zeroing memory buffers	Mai Lavelle
	Transferring memory to the device was very slow and there's really no need when only zeroing a buffer.
2017-03-08	Cycles: Add more atomic operations	Mai Lavelle

2017-03-08	Cycles: Expose passes size to device tasks	Mai Lavelle
	This is needed so devices can know the size of a tile buffer before any tiles are acquired.
2017-03-08	Cycles: Allow device_memory to be used directly	Mai Lavelle
	This is useful for when theres no host side memory attched to the buffer
2017-03-06	Cycles: Fix strict -Wpedantic warnings with GCC	Sergey Sharybin
	Patch by Stefan Werner, thanks!
2017-03-06	FFmpeg: Update for the deprecated API in 3.2.x	Sergey Sharybin
	Should be no functional changes.
2017-03-05	Fix T50843: Pitched Audio renders incorrectly in VSE	Jörg Müller
	There was a bug in the intended code behaviour to always seek with a pitch of 1.0 regardless of pitch/pitch animation/doppler effects. Check the bug report for a more detailed explanation of problems concerning pitch and seeking.
2017-03-03	Cycles: Fix possibly uninitialized variable	Sergey Sharybin
	Hopefully this was a reason of randomly disappearing textures in our renders.
2017-03-01	Cleanup: Use explicit unsigned int in atomics	Sergey Sharybin

2017-02-28	Cycles: Fix division by zero in volume code which was producing -nan	Sergey Sharybin

2017-02-28	Cleanup: Grey --> Gray	Aaron Carlisle

2017-02-27	Cycles: Fix compilation error on vanilla Ubuntu 16.10	Sergey Sharybin
	Patch by @swerner, thanks!
2017-02-27	Fix T50634: Hair Primitive as Triangles + Hair shader with a texture = crash	Sergey Sharybin
	Attributes were not resized after pushing new triangles to the mesh.
2017-02-27	Fix part of T50634: Hair Primitive as Triangles + Hair shader with a texture ↵	Sergey Sharybin
	= crash Wrong formula was used to calculate needed verts and tris to be reserved.
2017-02-27	Cycles: Make it more clear values never changes by using const qualifier	Sergey Sharybin

2017-02-27	Cycles: Forgot this in previous commit	Sergey Sharybin

2017-02-27	Cycles: Add more logs about what's going on in shader optimization	Sergey Sharybin

2017-02-27	Cycles: Experiment with replacing Sharp Glossy with GGX when Filter Glossy ↵	Sergey Sharybin
	is used The idea is to make it simpler to remove noise from scenes when some prop uses Sharp glossy closure and causes noise in certain cases. Previously Sharp Glossy was not affected by Filter Glossy at all, which was quite confusing. Here is a file which demonstrates the issue: {F417797} After applying the patch all the noise from the scene is gone. This change also solves fireflies reported in T50700. Reviewers: brecht, lukasstockner97 Differential Revision: https://developer.blender.org/D2416
2017-02-25	Fix T50698: Cycles baking artifacts with transparent surfaces.	Brecht Van Lommel

2017-02-25	Fix build error with macOS / clang / c++11.	Brecht Van Lommel

2017-02-24	Cycles: Fix compilation warning with CUDA on OSX	Sergey Sharybin

2017-02-23	Cycles: Fix compilation error on 32bit Linux	Sergey Sharybin

2017-02-23	Cycles: Fix wrong render results with texture limit and half-float textures	Sergey Sharybin

2017-02-23	Cycles: Add utility function to convert float to half	Sergey Sharybin
	handles overflow and underflow, but not NaN/inf.
2017-02-23	Fix T50748: Render Time incorrect when refreshing rendered preview in GPU mode	Sergey Sharybin

2017-02-22	Cycles: Fix shading with autosmooth and custom normals	Sergey Sharybin
	New logic of split_faces was leaving mesh in a proper state from Blender's point of view, but Cycles wanted loop normals to be "flushed" to vertex normals. Now we do such a flush from Cycles side again, so we don't leave bad meshes behind. Thanks Bastien for assistance here!
2017-02-22	Cycles: Additionally report all OpenCL cflags	Sergey Sharybin
	This way we can control exact spaces and such added to the cflags which is crucial to troubleshoot certain drivers.
2017-02-21	Cycles: Speedup by avoiding extra calculations in noise texture when unneeded	Mai Lavelle
	Noise texture is now faster when the color socket is unused. Potential for speedup spotted by @nutel. Some performance results: Render Time Before After Difference Gooseberry benchmark 47:51.34 45:55.57 -4% Koro 12:24.92 12:18.46 -0.8% Simple cube (Color socket) 48.53 48.72 +0.3% Simple cube (Fac socket) 48.74 32.78 -32.7% Goethe displacement 1:21.18 1:08.47 -15.6% Cycles brick displacement 3:02.38 2:16.76 -25.0% Large displacement scene 23:54.12 20:09.62 -15.6% Reviewed By: sergey Differential Revision: https://developer.blender.org/D2513
2017-02-20	Fix T50718: Regression: Split Normals Render Problem with Cycles	Sergey Sharybin
	The issue seems to be caused by vertex normal being re-calculated to something else than loop normal, which also caused wrong loop normals after re-calculation. For now issue is solved by preserving CD_NORMAL for loops after split_faces() is finished, so render engine can access original proper value.