XTra.KrazzY
|
fb60b10c7b
|
OpenCL: Implemented I4 and corrected IA8 (huge thanks to Orphis for fixes, implementation and optimization)
git-svn-id: https://dolphin-emu.googlecode.com/svn/trunk@4417 8ced0084-cf51-0410-be5f-012b33b47a6e
|
2009-10-13 21:51:27 +00:00 |
XTra.KrazzY
|
9da9346274
|
OpenCL: Kernels are now more general...
git-svn-id: https://dolphin-emu.googlecode.com/svn/trunk@4402 8ced0084-cf51-0410-be5f-012b33b47a6e
|
2009-10-11 22:21:36 +00:00 |
XTra.KrazzY
|
fc1c77b60d
|
OpenCL: Set up ground for more texture formats. Implemented the IA8 format (still buggy, no transparency)
git-svn-id: https://dolphin-emu.googlecode.com/svn/trunk@4401 8ced0084-cf51-0410-be5f-012b33b47a6e
|
2009-10-11 21:14:02 +00:00 |
XTra.KrazzY
|
449abdb01d
|
OpenCL: Solved the IA4 case (some byteswapping was necessary, hope there's a vectorized function for that). Some \n fixing I guess
git-svn-id: https://dolphin-emu.googlecode.com/svn/trunk@4399 8ced0084-cf51-0410-be5f-012b33b47a6e
|
2009-10-11 20:03:55 +00:00 |
XTra.KrazzY
|
e6f20f07f5
|
OpenCL: Vectorized I8 texture decoding and added a (too intense?) IA4 decoding method.
Not sure what's wrong with IA4 there...
git-svn-id: https://dolphin-emu.googlecode.com/svn/trunk@4382 8ced0084-cf51-0410-be5f-012b33b47a6e
|
2009-10-07 22:51:32 +00:00 |
XTra.KrazzY
|
f52851ab9f
|
OpenCL: Texture decoding is now two dimensional
git-svn-id: https://dolphin-emu.googlecode.com/svn/trunk@4381 8ced0084-cf51-0410-be5f-012b33b47a6e
|
2009-10-07 21:51:40 +00:00 |
XTra.KrazzY
|
98db1a0ca3
|
OpenCL: More windows support and some unfinished work
git-svn-id: https://dolphin-emu.googlecode.com/svn/trunk@4380 8ced0084-cf51-0410-be5f-012b33b47a6e
|
2009-10-07 21:30:36 +00:00 |
Sonicadvance1
|
5523f28b51
|
Better TF_I8 decoding via OCL. Can still get way better though. Seems XFB speed boost was a fluke
git-svn-id: https://dolphin-emu.googlecode.com/svn/trunk@4379 8ced0084-cf51-0410-be5f-012b33b47a6e
|
2009-10-07 21:10:39 +00:00 |
Sonicadvance1
|
16c6ec6e7e
|
Woops, that is why getting work group size failed
git-svn-id: https://dolphin-emu.googlecode.com/svn/trunk@4372 8ced0084-cf51-0410-be5f-012b33b47a6e
|
2009-10-07 07:44:35 +00:00 |
Sonicadvance1
|
059970e971
|
Fix XFB converting, silly mistake. Disable Texture converting ATM since it fails. And if can't read work group size, fall to 64, which is what my 8600GTS does. Tested XFB in Melee intro, got around 10FPS faster tPS faster then with CPU side
git-svn-id: https://dolphin-emu.googlecode.com/svn/trunk@4371 8ced0084-cf51-0410-be5f-012b33b47a6e
|
2009-10-07 04:00:29 +00:00 |
Sonicadvance1
|
0146f3f58f
|
More OpenCL work, got XFB converting which needs more optimization. I haven't checked for FPS changes. My desktop isn't the best to test on anyway (Phenom 1, 32 stream processors). The package check doesn't work for me, so I just checked true if you compile with opencl=true. Requires a bit of cleanup still
git-svn-id: https://dolphin-emu.googlecode.com/svn/trunk@4369 8ced0084-cf51-0410-be5f-012b33b47a6e
|
2009-10-07 02:48:21 +00:00 |
XTra.KrazzY
|
c6f0335b62
|
OpenCL: More work on centralization (need SCons reflection)
git-svn-id: https://dolphin-emu.googlecode.com/svn/trunk@4351 8ced0084-cf51-0410-be5f-012b33b47a6e
|
2009-09-30 21:01:34 +00:00 |
XTra.KrazzY
|
4499444a86
|
OpenGL and OpenCL are NOT related! (Committed here are some scons fixes)
git-svn-id: https://dolphin-emu.googlecode.com/svn/trunk@4349 8ced0084-cf51-0410-be5f-012b33b47a6e
|
2009-09-30 12:58:02 +00:00 |
nakeee
|
ffe12ec752
|
OpenCL: added opencl=true option to scons and fixed the case on the way.
Add ifdef for osx ocl include.
git-svn-id: https://dolphin-emu.googlecode.com/svn/trunk@4348 8ced0084-cf51-0410-be5f-012b33b47a6e
|
2009-09-30 06:49:08 +00:00 |
XTra.KrazzY
|
aa1fd81e37
|
OpenCL: A bit more work.
git-svn-id: https://dolphin-emu.googlecode.com/svn/trunk@4347 8ced0084-cf51-0410-be5f-012b33b47a6e
|
2009-09-30 00:39:32 +00:00 |
Sonicadvance1
|
fb258e3ca5
|
Base of OpenCL code
git-svn-id: https://dolphin-emu.googlecode.com/svn/trunk@4346 8ced0084-cf51-0410-be5f-012b33b47a6e
|
2009-09-30 00:28:27 +00:00 |