rodolfoosvaldobogado
53b62ab169
This should fix the bugs introduced by my last commit please retest the games that where showing graphical glitches.
2012-11-10 11:37:08 -03:00
NeoBrainX
6f21f5eb34
Video_Software: Implement texture preloading
2012-11-03 15:41:41 +01:00
Ryan Houdek
fb92c338af
Bit of cleanup and fixing of one issue that was noticeable in SMS with Mario's shadow.
2012-10-27 22:50:06 -05:00
Ryan Houdek
dfb3c44d1a
Remove glMatrixMode and glLoadIdentity calls in Render.cpp which do absolutely nothing.
2012-10-27 18:27:16 -05:00
Ryan Houdek
8c1d104ca3
Change EFB copy location rectangles over to a Vertex array
2012-10-27 18:10:31 -05:00
Ryan Houdek
c1fd640964
Rebase immediate-removal on master
2012-10-27 17:22:48 -05:00
rodolfoosvaldobogado
ee72852491
implement some code to reduce the amounts of calls to setup vertex format, in d3d9 it gives no noticeable speedup, in opengl it still does not work right.
...
thanks to neobrain for the idea
2012-10-26 23:18:09 -03:00
rodolfoosvaldobogado
eaa1ea71c1
Implement the new buffer approach in opengl. sadly in my machine it gives my only 2 more fps and if your hardware does not support ARB_map_buffer_range is even slower than plain vertex arrays.
...
change naming in all the backends vertex managers to make more easy to continue with the merge an some future improvements.
please test this as i'm interested in knowing the performance in linux and windows with the different hardware platforms.
2012-10-26 11:34:02 -03:00
rodolfoosvaldobogado
72cb1734ee
Some code cleaning for my last commit.
...
the amount and size of the buffer is now changed to "new hardware" frienly values and will fall back to the right values if hardware does not support them.
my next commit will be to a branch, with my ogl work.
2012-10-22 19:37:26 -03:00
NeoBrainX
52963584f0
[bugfix] Make sure to specify the correct max lod level.
2012-10-20 21:07:03 +02:00
NeoBrainX
9d5fc3d7f0
[bugfix] Properly decode lod_bias.
2012-10-20 21:07:03 +02:00
NeoBrainX
6c8a22de24
[bugfix] DX9::TextureCache: Use max_lod instead of min_lod where necessary.
2012-10-20 21:07:03 +02:00
NeoBrainX
0706050b74
[cleanup] Vastly clean up backend-specific sampler state logic.
2012-10-20 21:07:02 +02:00
NeoBrainX
069d949a57
[cleanup] TextureCache: Kill the autogen_mipmap parameter
2012-10-20 21:07:02 +02:00
rodolfoosvaldobogado@gmail.com
5230146c73
Hey, long time no commits :).
...
So to compensate lets bring back some speed to the emulation.
change a little the way the vertex are send to the gpu,
This first implementation changes dx9 a lot and dx11 a little to increase the parallelism between the cpu and gpu.
ogl: is my next step in ogl is a little more trickier so i have to take a little more time.
the original concept is Marcos idea, with my little touch to make it even more faster.
what to look for: SPEEEEEDDD :).
please test it a lot and let me know if you see any problem.
in dx9 the code is prepared to fall back to the previous implementation if your card does not support the amount of buffers needed.
So if you did not experience any speed gains you know where is the problem :).
for the ones with more experience and compression of the code please test changing the amount and size of the buffers to tune this for your specific machine.
The current values are the sweet spot for my machine.
All must Thanks Marcos, I hate him for giving good ideas when I'm full of work.
2012-10-20 10:22:15 -03:00
Ryan Houdek
57426ee726
Forced commit
2012-10-13 13:00:04 -05:00
Shawn Hoffman
1a8005d948
set ProgramShaderCache program format correctly.
2012-10-09 23:56:00 -05:00
Ryan Houdek
eb7a0c485a
More "stuff" for SS, also a small fix in the program shader cache cache setup."
2012-10-09 23:56:00 -05:00
Shawn Hoffman
210ecad15f
let us try normal c++ static init instead...
2012-10-09 23:56:00 -05:00
Ryan Houdek
03b09bed5d
Get the program binary type correctly or else ATI makes massive (~400MB) shader caches. Also, don't need the line in the PixelShaderGen.
2012-10-09 23:56:00 -05:00
Ryan Houdek
2e15440896
Add support for Dual source blending to older ATI cards that don't support 420pack but do support GL_ARB_blend_func_extended. This is more proper as well anyways.
2012-10-09 23:56:00 -05:00
Ryan Houdek
5085cebaf3
These checks aren't needed when using CG.
2012-10-09 23:55:59 -05:00
Shawn Hoffman
108722bed0
oops, forgot to make sure the gl program is actually free'd.
2012-10-09 23:54:18 -05:00
Ryan Houdek
e3854ded73
Woops, better not forget the ing
2012-10-09 23:54:18 -05:00
Ryan Houdek
4cd748bbec
Remove some warnings in ProgramShadercache, Was using wrong variable for checking dual source blending.
2012-10-09 23:54:18 -05:00
Shawn Hoffman
4c1fef8f98
refactor ProgramShaderCache::PCacheEntry
2012-10-09 23:54:18 -05:00
Shawn Hoffman
aaa405c973
Checking GLEW_VERSION_4_0 is superfluous since we check GL_ARB_get_program_binary, and it's a runtime variable anyways.
2012-10-09 23:54:18 -05:00
Shawn Hoffman
31a8424bcc
fix formatting uglies introduced in glsl-master branch
2012-10-09 23:54:17 -05:00
Ryan Houdek
c4e7a288e5
Add in the Windows fix.
2012-10-09 23:43:22 -05:00
Ryan Houdek
ee529b7125
Only use explicit attribute locations when we are supporting GLSL > 1.2 since we need in/out instead of attribute variable types. This was brought to my attention from MESA. MESA supports GL_ARB_explicit_attrib_location, but yets to support GLSL 1.3, so basically useless extension to MESA right now?
2012-10-09 23:43:22 -05:00
Ryan Houdek
5c486587a6
Write all shaders to disk on emulator stop instead of constantly. Also change pair from u64 to u32.
2012-10-09 23:43:22 -05:00
Ryan Houdek
4eb227bd4c
If user doesn't compile with glew 1.6, cut out program binaries so everything still compiles.
2012-10-09 23:43:22 -05:00
Ryan Houdek
189d12b61c
If CG isn't available, still compile and fallback on GLSL
2012-10-09 23:43:22 -05:00
Sonicadvance1
7c91476650
Missed a if, don't want people crashing now do we?
2012-10-09 23:42:41 -05:00
Ryan Houdek
9996f27120
Give OSX users more of a chance of supporting Single pass DSB in the future.
2012-10-09 23:42:41 -05:00
Ryan Houdek
d012c75005
Implement Program shaders cache. Seems to reduce a small amount of stuttering when F-Zero starts. Did it because I can :|
2012-10-09 23:42:41 -05:00
Ryan Houdek
ef1e157786
Go back to using glGetString with GL_EXTENSIONS because glew 1.7 is required to check for newer 4.2 extensions that way.
2012-10-09 23:42:40 -05:00
Ryan Houdek
ac77bbd47b
Move the GLSL extension checks over here so it is nicer.
2012-10-09 23:42:40 -05:00
Ryan Houdek
7cb3d86e31
Add a define back so we work without GLEW 1.6, Also I missed a line when I was fixing binding sampler locations last night when I was tired.
2012-10-09 23:42:40 -05:00
Ryan Houdek
1aad2e9e56
There we go, actually found the issue.
2012-10-09 23:42:40 -05:00
Ryan Houdek
6759ee701d
This thing fails when we don't support binding. Not 100% sure why I need to set this multiple times.
2012-10-09 23:42:40 -05:00
Ryan Houdek
66a245dd68
Remove this silliness
2012-10-09 23:42:40 -05:00
Ryan Houdek
9c09e1973a
Some shader programs don't come with Vertex shaders attached, don't try to bind a Vertex Uniform block then.
2012-10-09 23:42:40 -05:00
Ryan Houdek
f8d0c28e53
Set Sampler values at program make time instead of every frame. Fix an issue when The user had UBO support but not Binding support.
2012-10-09 23:42:40 -05:00
Ryan Houdek
ddd4360d6d
Only delete this buffer if we support it.
2012-10-09 23:42:39 -05:00
Ryan Houdek
7aad45658e
Add a GUI option to use GLSL shaders. Also fix a small typo.
2012-10-09 23:42:39 -05:00
Ryan Houdek
076c1a5aa7
Make sure not to try and bind UBO locations when it isn't supported
2012-10-09 23:41:49 -05:00
Ryan Houdek
ba12c0b4f5
Instead of querying the vertex attribute location. Let's bind it to where CG expects it to be as well. Was causing problems when we were trying to activate the components below and they weren't available. This fixes cubivore with GLSL shader. Also any other game that uses 3 normal pointers.
2012-10-09 23:41:48 -05:00
Jordan Woyak
d70726b035
glMapBuffer was slow, go back to glBufferSubData, single combined ps/vs ubo now
2012-10-09 23:41:48 -05:00
Jordan Woyak
d9117ab6a1
try combining vs/ps ubo
2012-10-09 23:41:48 -05:00
