kannonier8/blender
1
0
Fork 0
forked from bartvdbraak/blender
Code Pull Requests Activity
d8509b349d
blender/intern/cycles/kernel
History
Brecht Van Lommel 6da6f8d33f Cycles: CUDA faster rendering of small tiles, using multiple samples like 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
2017-10-04 21:58:47 +02:00
..
bvh Cycles: disable fast math flags, only use a subset. 2017-09-08 15:12:37 +02:00
closure Code cleanup: refactor BSSRDF closure sampling, for next commit. 2017-09-20 19:38:08 +02:00
filter Code refactor: use split variance calculation for mega kernels too. 2017-10-04 21:11:14 +02:00
geom Fix T52660: CUDA volume texture rendering not working on Fermi GPUs. 2017-09-06 18:12:45 +02:00
kernels Cycles: CUDA faster rendering of small tiles, using multiple samples like OpenCL. 2017-10-04 21:58:47 +02:00
osl Fix Cycles CUDA transparent shadow error after recent fix in c22b52c. 2017-08-24 03:43:02 +02:00
shaders Fix T52152: allow zero roughness for Cycles principled BSDF, don't clamp. 2017-07-22 23:58:51 +02:00
split Code refactor: zero render buffers outside of kernel. 2017-10-04 21:11:14 +02:00
svm Fix T52645, T52645: AMD OpenCL compiler crash with recent drivers. 2017-10-04 21:00:46 +02:00
CMakeLists.txt CMake: support CUDA 9 toolkit, and automatically disable sm_2x binaries. 2017-10-01 14:14:53 +02:00
kernel_accumulate.h Cycles: reduce subsurface stack memory usage. 2017-09-28 15:18:43 +02:00
kernel_bake.h Fix use of uninitialized memory in Cycles normal baking. 2017-10-04 21:11:14 +02:00
kernel_camera.h Cycles: Remove ccl_fetch and SOA 2017-03-08 00:52:41 -05:00
kernel_compat_cpu.h Cycles: Use falltrhough attribute to help catching missing break statements 2017-05-24 17:23:54 +02:00
kernel_compat_cuda.h Cycles: Attempt to work around compilation of sm_20 and sm_21 2017-09-08 18:37:54 +02:00
kernel_compat_opencl.h Cycles: Add utility macro ccl_ref 2017-08-08 15:27:25 +02:00
kernel_differential.h Cycles: OpenCL kernel split 2015-05-09 19:52:40 +05:00
kernel_emission.h Cycles: slightly improve BSDF sample stratification for path tracing. 2017-09-20 19:38:08 +02:00
kernel_film.h Cycles: Use native saturate function for CUDA 2015-04-28 00:38:32 +05:00
kernel_globals.h Cycles: Remove ulong usage 2017-08-09 14:08:58 +02:00
kernel_image_opencl.h Cycles: Pack kernel textures into buffers for OpenCL 2017-08-08 07:12:04 -04:00
kernel_jitter.h Cycles: Use more stable version of integer square root function 2017-05-09 17:07:17 +02:00
kernel_light.h Code cleanup: avoid used of uninitialized value in case of precision issue. 2017-09-13 15:24:14 +02:00
kernel_math.h Cycles: Make all #include statements relative to cycles source directory 2017-03-29 13:41:11 +02:00
kernel_montecarlo.h Cycles code refactor: minor refactoring and comments for volume code. 2014-03-29 13:03:49 +01:00
kernel_passes.h Cycles: CUDA faster rendering of small tiles, using multiple samples like OpenCL. 2017-10-04 21:58:47 +02:00
kernel_path_branched.h Code refactor: zero render buffers outside of kernel. 2017-10-04 21:11:14 +02:00
kernel_path_common.h Code refactor: remove rng_state buffer and compute hash on the fly. 2017-10-04 21:11:14 +02:00
kernel_path_state.h Code cleanup: remove hack to avoid seeing transparent objects in noise. 2017-09-20 19:38:08 +02:00
kernel_path_subsurface.h Cycles: reduce subsurface stack memory usage. 2017-09-28 15:18:43 +02:00
kernel_path_surface.h Cycles: reduce subsurface stack memory usage. 2017-09-28 15:18:43 +02:00
kernel_path_volume.h Cycles: reduce subsurface stack memory usage. 2017-09-28 15:18:43 +02:00
kernel_path.h Code refactor: zero render buffers outside of kernel. 2017-10-04 21:11:14 +02:00
kernel_projection.h Cycles: Implement denoising option for reducing noise in the rendered image 2017-05-07 14:40:58 +02:00
kernel_queues.h Cycles: Add function to dequeue a ray 2017-06-10 03:51:18 -04:00
kernel_random.h Code refactor: remove rng_state buffer and compute hash on the fly. 2017-10-04 21:11:14 +02:00
kernel_shader.h Cycles: Fix compilation error of megakernel on NVidia device 2017-09-23 17:03:02 +05:00
kernel_shadow.h Cycles: Fix compilation error of OpenCL megakernel on Apple 2017-09-23 17:07:19 +05:00
kernel_subsurface.h Cycles: slightly improve BSDF sample stratification for path tracing. 2017-09-20 19:38:08 +02:00
kernel_textures.h Fix T52660: CUDA volume texture rendering not working on Fermi GPUs. 2017-09-06 18:12:45 +02:00
kernel_types.h Code refactor: remove rng_state buffer and compute hash on the fly. 2017-10-04 21:11:14 +02:00
kernel_volume.h Cycles: slightly improve BSDF sample stratification for path tracing. 2017-09-20 19:38:08 +02:00
kernel_work_stealing.h Code refactor: add WorkTile struct for passing work to kernel. 2017-10-04 21:11:14 +02:00
kernel.h Cycles: Implement denoising option for reducing noise in the rendered image 2017-05-07 14:40:58 +02:00