thirdparty/ppsspp
2
0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions
19956 commits 39 branches 0 tags 279 MiB
Commit graph

107 commits

Author SHA1 Message Date
Unknown W. Brackets
e8c8f19aec Move framebuf attachment into central code. 
It doesn't look like this will be different, and there's a bug in it.
 2016-03-26 12:30:32 -07:00
Henrik Rydgard
bd17351ef9 Texcache: Minor refactor/cleanup 2016-03-22 23:56:55 +01:00
Unknown W. Brackets
add506a0d3 Cleanup differences in aniso handling. 
Some places were not handling it as a power of two.
 2016-03-20 19:31:02 +01:00
Henrik Rydgard
dced84c9fc Work on texture support, cube.elf works correctly. Delete unused code. 2016-03-20 19:30:11 +01:00
Unknown W. Brackets
e30ab95179 d3d9: Fix maxSeenV when not known. 
And just make the code reused between both.  Fixes #8478.
 2016-01-22 18:48:54 -08:00
Unknown W. Brackets
909d477719 Support CLUTs at an x offset. 
Used by Kurohyo 2.  Highly unlikely to be a mis-estimate within stride.
 2016-01-05 00:02:58 -08:00
Unknown W. Brackets
28a07c70c6 Explicitly download rendered cluts. 
This avoids triggering logic that tries to get the sizing right, or
optimize frequent copies.  CLUTs often get estimated wrong, so it's better
to copy just the correct range, always.
 2016-01-04 21:29:03 -08:00
Unknown W. Brackets
36c8b043f3 d3d9: Fix CLUT4 optimization color creation. 
Fixes #8385.
 2016-01-04 07:36:47 -08:00
Unknown W. Brackets
6cc69ed25f Factor out texture shader application. 
This allows it to be reused between depal and indexed textures (second not
implemented yet.)
 2016-01-03 23:50:49 -08:00
Unknown W. Brackets
14b350a5fe d3d9: Fix render-to-texture not working. 
Broken by a copy/paste mistake in #7870 form OpenGL.
 2016-01-03 23:44:43 -08:00
Unknown W. Brackets
e427d09628 Determine scaleFactor once per frame. 
This way, for auto, we can properly treat it like scaleFactor = 1 when the
resolution is small.
 2016-01-03 23:06:15 -08:00
Unknown W. Brackets
2494176919 Remove incorrect CLUT4 optimization. 2016-01-02 13:20:30 -08:00
Unknown W. Brackets
69daa75228 Move more common code around in texcache. 2015-12-31 11:41:02 -08:00
Unknown W. Brackets
a45f05c598 Move NotifyFramebuffer and LoadClut to common code. 2015-12-31 11:13:39 -08:00
Unknown W. Brackets
7a864f5aa1 Be forgiving when maxSeenV increases. 
This still allows us to optimize < 512 heights, but at the same time,
doesn't hit spritesheets as hard and avoids the scaling penalty.
 2015-12-30 17:15:50 -08:00
Unknown W. Brackets
d18164bd4d d3d9: Clean up some scaling handling. 2015-12-30 16:54:25 -08:00
Unknown W. Brackets
1339f6b1f3 Report formats on usage of rendered clut. 2015-11-26 14:31:46 -08:00
Unknown W. Brackets
3965cba187 Disable clut render for now as a "slow effect". 
Also, add reporting.
 2015-11-26 13:42:18 -08:00
Unknown W. Brackets
8e2557bafe Support render-to-clut, at least in some cases. 
This is pretty much only tested with Brave Story.  See #6754.

There may be other cases which are not handled yet.
 2015-11-26 13:31:17 -08:00
Henrik Rydgard
3acdd4e237 We had two global types called FBO, which is not ok. Fix that. 
This caused some very strange behaviour in the MSVC debugger when looking at this stuff.
 2015-11-14 13:24:51 +01:00
Henrik Rydgard
7e3cd987cd Fixes to D3D viewport management. For some reason, still having strange driver-bug-ish problems in unbuffered... 2015-11-12 14:47:43 +01:00
Henrik Rydgard
6b2862284b Minor cleanups 2015-11-02 20:09:59 +01:00
Henrik Rydgard
069a7b0d2f Move the viewport to the right place. 2015-11-02 20:08:00 +01:00
Henrik Rydgard
f4248cb550 Remove a whole lot of flipping special cases 2015-11-02 20:07:30 +01:00
Unknown W. Brackets
dff4aeb30f Correct uv bound pointers in arm vertexjit. 2015-09-13 14:55:10 -07:00
Unknown W. Brackets
93e9d6cd56 Respect framebuffer offset in fb copies. 
We need to offset the copy coordinates also, or we'll copy the wrong
thing.
 2015-09-13 11:35:15 -07:00
Unknown W. Brackets
62de281e35 Use texture UV range to optimize framebuf copies. 2015-09-13 11:34:52 -07:00
Unknown W. Brackets
f4df7f076e Hash less of 512 tall textures when possible. 
This allows xBRZ to work better on such textures, and also may reupload
them less frequently.
 2015-09-13 11:15:57 -07:00
Unknown W. Brackets
52c3fc221b Also split DirectX9's texture cache. 2015-09-13 09:36:17 -07:00
Unknown W. Brackets
71afaffc20 Pack texture cache entries just a bit tighter. 2015-09-13 06:43:47 -07:00
Unknown W. Brackets
7d63a10fdd Centralize texture cache entry structures. 2015-09-13 06:43:37 -07:00
Unknown W. Brackets
e1df8fc45f Centralize sampling params and test optimizations. 2015-09-13 06:38:07 -07:00
Henrik Rydgard
1c68069083 refactor: Get rid of gstate from depalshadercache 2015-07-29 14:28:51 +02:00
Henrik Rydgard
53f062b7b2 refactor: Remove some uses of gstate from SplineCommon and a fex texcache functions 2015-07-29 11:40:45 +02:00
Henrik Rydgard
e857bd9a1c Minor code simplification in texcache 2015-07-29 10:59:02 +02:00
Unknown W. Brackets
c75010931c Use SSE in CheckAlpha scanning. 2015-05-24 22:55:43 -07:00
Unknown W. Brackets
4fa50a9a50 Reduce clut allocation size a bit. 
Really think this probably should only be 1024 bytes, but need to work out
some more details before shrinking all the way.
 2015-04-26 00:50:47 -07:00
Unknown W. Brackets
ab67c49ae9 Make sure we don't hash outside max bytes. 
If we've never even loaded that much, play it safe.
 2015-04-26 00:43:36 -07:00
Unknown W. Brackets
c8fc9b0bf3 Cleanup some incorrect handling of clut offset. 2015-04-26 00:31:00 -07:00
Henrik Rydgård
a4a10588ed Merge pull request #7697 from unknownbrackets/texcache 
Include texture size in the cache key
 2015-04-20 00:08:44 +02:00
Unknown W. Brackets
b726a414d7 Log a reason when invalidating textures. 
Makes debugging easier.
 2015-04-18 12:42:13 -07:00
Unknown W. Brackets
58563324bd Include texture size in cache key. 
This makes Tales of Destiny 2's towns significantly faster.  It may
however cause us to keep textures around for longer - but we still account
for them in our metrics and invalidation.
 2015-04-18 12:42:13 -07:00
Unknown W. Brackets
ce9f404bef Unswizzle when reading from the swizzled mirror. 2015-04-18 12:39:04 -07:00
Unknown W. Brackets
1e8f2c2630 Allow clut load outside valid mem, fill with zero. 
And only for the parts outside memory.
 2015-04-12 22:38:17 -07:00
Unknown W. Brackets
10626e356d Avoid a potential divide by zero. 2015-04-08 11:57:59 -07:00
Unknown W. Brackets
2450724be2 Make Memory::Memcpy() execute memchecks directly. 
This makes it easier to handle breakpoints in HLE.
 2015-04-05 18:09:20 -07:00
Unknown W. Brackets
7ee9112425 Avoid constant reload of frequently changed tex. 
If we detect that it's frequently changing, this will cause it to spiral
into an infinite loop of staying frequently changing and reloading to try
to scale.
 2015-03-15 09:10:09 -07:00
Unknown W. Brackets
3bc954778c dx9: Explicitly restore viewport after depal. 
Otherwise, it doesn't get restored.
 2015-03-01 11:28:06 -08:00
Unknown W. Brackets
bda6f0a4b8 dx9: Reset more state before depal draw. 
Was alpha blending before.
 2015-03-01 11:28:05 -08:00
Henrik Rydgard
869f74d1b0 Depal: Apply the half-pixel offset that looked the least wrong. Still not right. 2015-03-01 11:28:04 -08:00