Sergey Sharybin 42aeb608e7 Cycles: Implement AVX2 version of triangle_intersect ... This commit basically vectorizes existing code using AVX2 instructions (without modifying algorithm itself). This gives quite nice speedups: BMW: -8% Classroom: -5% Cat: -5% Koro: +1% Barcelona: -8% That's on Linux machine, reported performance improvement on Windows goes up to 20%. Not currently sure why Koro is somewhat slower because it mainly uses curve intersection tests, could be a time noise? Or osmething with the cache utilization perhaps? In any case speedup in other scenes makes me thinking that current state is acceptable for initial implementation. This is again inspired by Maxym Dmytrychenko.		2016-10-12 14:11:55 +02:00
build_files	install_deps.sh: remove libglewmx from deb for now, the package is broken in deb testing...	2016-10-10 19:29:55 +02:00
doc	API doc update script: Fix generated zipfile name, was broken in 'release' case...	2016-09-13 08:39:01 +02:00
extern	CUEW: Use latest upstream version	2016-09-30 10:36:11 +02:00
intern	Cycles: Implement AVX2 version of triangle_intersect	2016-10-12 14:11:55 +02:00
release	Fix objects added via py being on the wrong layer when viewport is decoupled from scene	2016-10-11 15:35:14 +02:00
source	Fix T49622: Grease pencil not rendering out of VSE	2016-10-12 10:01:30 +02:00
tests	Fix undefined variable on errors in Cycles ctests.	2016-09-10 23:04:14 +02:00
.arcconfig	Use HTTPS protocol for arc	2013-12-24 22:57:27 +06:00
.gitignore	Update .gitignore to some more generated/non-git-stored files.	2016-08-21 22:40:57 +02:00
.gitmodules	Add developer tools submodule	2016-07-15 01:28:00 +10:00
CMakeLists.txt	CMake changes for new macOS target 10.9 / libc++ libraries.	2016-10-09 01:54:29 +02:00
COPYING	== docs ==	2010-10-13 14:44:22 +00:00
GNUmakefile	Cleanup: line length	2016-07-15 01:26:26 +10:00
make.bat	[msvc] make.bat, no need to set the cuda path at all, cmake will figure it out on it's own.	2016-10-04 13:30:38 -06:00