Commit Graph

56 Commits

Author SHA1 Message Date
Gregory Hainaut ad63c13cd1 gsdx sw: Update comment on the GCC bug
Add link to bugzilla. Bug will be fixed in GCC7

Close #1881
2017-04-04 19:04:21 +02:00
Gregory Hainaut df32564bef gsdx: workaround AVX2 generated code by GCC
See commit for the full details.

v2: use a direct access to the union field instead of extract32
It gives us both optimal and working code.
2017-04-01 14:05:40 +02:00
Gregory Hainaut f862f5be03 gsdx sw: use accurate division for the C reference implementation
Speed isn't important here. It would allow to compare the rendering with
the JIT implementation. If it is necessary we could an option for the JIT.
2017-03-10 20:45:14 +01:00
Gregory Hainaut 8431299b92 gsdx sw: port code to the new constant object 2016-11-24 23:03:26 +01:00
Gregory Hainaut 3b5bc9c38d gsdx sw: create a constant buffer
* Use POD type to avoid SSE/AVX compilation dependency
* global object to reduce cache miss
* dynamically object so give a chance to allocate below 2GB (allow x64
  optimization)
2016-11-24 23:03:26 +01:00
Gregory Hainaut aa4b2d9f3a gsdx sw: wrap GS memory
Cost ought to remain small. Worst case is 2 extra "and" operation by group of pixels in scanline renderer

I think PixelAddressN functions are mostly call in the init.
2016-11-01 13:33:11 +01:00
gabest11 d631030608 gsdx: fixed some of the renames where it made no sense, also added a commend about lupin 3rd. 2015-08-04 03:36:42 +02:00
gabest11 a1a842b07f gsdx: date/datm fix for 16-bit frame buffer in sw rendering mode 2015-08-02 20:21:49 +02:00
Gregory Hainaut 28bb64aae8 gsdx: sed/dr/dRect/ 2015-05-15 20:49:32 +02:00
Gregory Hainaut 6a9e425308 gsdx: sed /dt/dTex/ 2015-05-15 20:44:15 +02:00
gabest11 ba1e522bbb Less opencl bugs, some games are actually playable now, there are still texture errors. 2014-12-02 00:16:34 +01:00
gabest11@gmail.com 4dd898c754 GSdx: Found some 16-bit integer overflow in GSState::VertexKick, some triangles could have been removed by the scissor test.
git-svn-id: http://pcsx2.googlecode.com/svn/trunk@5680 96395faa-99c1-11dd-bbfe-3dabce05a288
2013-06-24 01:11:00 +00:00
gabest11@gmail.com 20d99ae9fc GSdx: vs2010 fix and minor changes
git-svn-id: http://pcsx2.googlecode.com/svn/trunk@5678 96395faa-99c1-11dd-bbfe-3dabce05a288
2013-06-23 10:46:24 +00:00
gabest11@gmail.com d20bd4f86a GSdx: The sw renderer now uses avx2, not much faster though, +10% maybe, if the game is not EE limited. I'm not sure if haswell has that much better sse execution (load/store units doubled for example), or the avx2 code is not fully optimized yet.
git-svn-id: http://pcsx2.googlecode.com/svn/trunk@5677 96395faa-99c1-11dd-bbfe-3dabce05a288
2013-06-20 05:07:52 +00:00
gabest11@gmail.com 43c7235cc3 GSdx: Added a vector class for AVX2 and upgraded a few functions to use it, still a lot to do. New project config AVX2, _M_SSE = 0x501, only use it if you have haswell. Might break GCC or older MSVC, please test it.
git-svn-id: http://pcsx2.googlecode.com/svn/trunk@5652 96395faa-99c1-11dd-bbfe-3dabce05a288
2013-06-10 15:56:11 +00:00
gabest11@gmail.com d27b606e5c GSdx:
- changed the KH2 fix in GetTextureMinMax to my taste, should give the same results, when the used texture rect is to the left/above the clamped area, it returns [min, min+1], and [max-1, max] for right/below
- m_mem.m_clut.Read32 was returned to its original place from GetAlphaMinMax
- UserHacks_WildHack was moved up to GSState, special UV handlers are only used when this setting is active
- updated xbyak to the latest available (avx2 seems incomplete, the 256-bit promoted old instructions are missing)
- changed vtune's include path to the 2013 edition

Some other not yet commited changes from a year ago:
- WriteImageX overflow crash-fix
- moved colclamp after dithering (sw mode), it makes more sense, no visible changes noticed
- Gif_Tag::analyzeTag(), there was a conditional I didn't like, split the loop into two parts

git-svn-id: http://pcsx2.googlecode.com/svn/trunk@5649 96395faa-99c1-11dd-bbfe-3dabce05a288
2013-06-06 11:36:01 +00:00
gregory.hainaut e06484adb8 gsdx: update the copyrigh address thank to sed
git-svn-id: http://pcsx2.googlecode.com/svn/trunk@5412 96395faa-99c1-11dd-bbfe-3dabce05a288
2012-09-09 18:16:11 +00:00
gabest11 67ef781116 GSdx: this may fix silent hill shadows and mister mosquito intro blur, also reduced texture cache keep-alive time from 30 to 10 frames and found two memory leaks, killzone can run a few seconds longer before crashing, I think there is something in pcsx2 allocating too much memory.
git-svn-id: http://pcsx2.googlecode.com/svn/trunk@5096 96395faa-99c1-11dd-bbfe-3dabce05a288
2012-02-12 17:56:06 +00:00
gabest11 da7324dbf2 Not transferring unused vif registers to the MTVU thread can save at least half of the ring buffer space. The whole set is about 400 bytes, including padding, but I could find references to only 6 regs.
git-svn-id: http://pcsx2.googlecode.com/svn/trunk@5092 96395faa-99c1-11dd-bbfe-3dabce05a288
2012-02-10 07:03:55 +00:00
gabest11 19be605150 GSdx: this should fix xp/wine crashing when extrathreads > 0, and added a sprite drawing shortcut, hopefully won't break anything.
git-svn-id: http://pcsx2.googlecode.com/svn/trunk@5089 96395faa-99c1-11dd-bbfe-3dabce05a288
2012-02-08 16:57:14 +00:00
gabest11 fcc09c2cca GSdx: fixing two different bugs of the sw renderer, addressing outside the texture in region wrap mode (skygunner), and little gaps in shadows and other random places (dq8, rogue galaxy, okami).
git-svn-id: http://pcsx2.googlecode.com/svn/trunk@5085 96395faa-99c1-11dd-bbfe-3dabce05a288
2012-01-29 10:12:20 +00:00
gabest11 bd12a8b7b3 GSdx: silent hill shattered memories purple street light fix, not block aligned filled rects were not correctly masked, may also fix games which store texture data in the alpha channel but use 24-bit target buffers (or just mask it in every time) and draw a lot of those unaligned rects. (tested many .gs dumps, very rare combination)
git-svn-id: http://pcsx2.googlecode.com/svn/trunk@5084 96395faa-99c1-11dd-bbfe-3dabce05a288
2012-01-28 10:07:17 +00:00
gabest11 481f1fdda2 GSdx: it's hard to keep track of the leftover vertices properly, a bit of sps was still possible, psx sprites were fixed too
git-svn-id: http://pcsx2.googlecode.com/svn/trunk@5065 96395faa-99c1-11dd-bbfe-3dabce05a288
2012-01-09 08:41:33 +00:00
gabest11 2eec75c2ae GSdx: sps fixed, some code clean up and optimization, ps2 logo still broken in hw mode, I'll check it later
git-svn-id: http://pcsx2.googlecode.com/svn/trunk@5062 96395faa-99c1-11dd-bbfe-3dabce05a288
2012-01-08 17:10:00 +00:00
gabest11 f68f007f00 GSdx: the promised index buffer update, needed a lot of changes, expect bugs in the next dozen revisions.
git-svn-id: http://pcsx2.googlecode.com/svn/trunk@5045 96395faa-99c1-11dd-bbfe-3dabce05a288
2012-01-05 02:40:24 +00:00
gabest11 2421c68bee GSdx: Saving the conditional var update (vista or better) before I try a new idea again. That Sync() call is wasting too much time, if there was only one queue then the main thread could also grab and process elements instead of just waiting for the workers.
git-svn-id: http://pcsx2.googlecode.com/svn/trunk@5005 96395faa-99c1-11dd-bbfe-3dabce05a288
2011-12-22 01:48:16 +00:00
gabest11 f318e84aca GSdx: Better multi-threading for the sw renderer. Threads must be synchronized lot less, 1/10th in average, can run parallel longer and uses more cpu (bit more empty spinning, too). There could be some new bugs, as usual.
git-svn-id: http://pcsx2.googlecode.com/svn/trunk@4992 96395faa-99c1-11dd-bbfe-3dabce05a288
2011-12-18 08:13:20 +00:00
gabest11 4b77052d21 GSdx: just saving minor changes.
git-svn-id: http://pcsx2.googlecode.com/svn/trunk@4991 96395faa-99c1-11dd-bbfe-3dabce05a288
2011-12-16 19:13:58 +00:00
gabest11 786f43a707 GSdx: more fixes to z-interpolation, THPS4 looks alright now.
git-svn-id: http://pcsx2.googlecode.com/svn/trunk@4977 96395faa-99c1-11dd-bbfe-3dabce05a288
2011-12-04 09:49:55 +00:00
gabest11 f1537f53a2 GSdx: Found where the bypassed int z was destroyed. (bug appeared in r4967)
git-svn-id: http://pcsx2.googlecode.com/svn/trunk@4975 96395faa-99c1-11dd-bbfe-3dabce05a288
2011-12-01 17:08:10 +00:00
gabest11 9d54677055 GSdx: re-implemented the drawing pipeline in c++, just for reference and easier debugging.
git-svn-id: http://pcsx2.googlecode.com/svn/trunk@4972 96395faa-99c1-11dd-bbfe-3dabce05a288
2011-11-25 23:48:59 +00:00
gabest11 4f6f53c188 GSdx: Just remembered triangles were occasionally converted to sprites, need to set t.w there as well.
git-svn-id: http://pcsx2.googlecode.com/svn/trunk@4967 96395faa-99c1-11dd-bbfe-3dabce05a288
2011-11-14 03:34:24 +00:00
gabest11 cc28241128 GSdx: just a small mipmapping optimization when lod is constant, it was already a TODO.
git-svn-id: http://pcsx2.googlecode.com/svn/trunk@4518 96395faa-99c1-11dd-bbfe-3dabce05a288
2011-04-01 12:36:21 +00:00
gabest11 b31634df8f GSdx: using mipmap levels (only per batch, no tri-linear) and a couple of small changes, including the stdcall fix for linux.
git-svn-id: http://pcsx2.googlecode.com/svn/trunk@4419 96395faa-99c1-11dd-bbfe-3dabce05a288
2011-03-12 22:10:58 +00:00
gabest11 3030166596 The core of GSdx is now compatible with intel's compiler on linux.
- GSWnd is not implemented, no config dialogs either
- no output, just the null device
- threading classes were not tested (my first experience with pthread)

git-svn-id: http://pcsx2.googlecode.com/svn/trunk@4315 96395faa-99c1-11dd-bbfe-3dabce05a288
2011-02-19 03:36:30 +00:00
gabest11 257d57ba52 Tweaked the rasterizer to be about 10% faster in multi-threaded mode (2 or 3 threads), still far from optimal.
git-svn-id: http://pcsx2.googlecode.com/svn/trunk@4308 96395faa-99c1-11dd-bbfe-3dabce05a288
2011-02-17 18:22:47 +00:00
gabest11 6f18c0dabe Trying to isolate the rasterizer step-by-step, for better multi-threading in the future.
git-svn-id: http://pcsx2.googlecode.com/svn/trunk@4305 96395faa-99c1-11dd-bbfe-3dabce05a288
2011-02-17 03:24:37 +00:00
gabest11 65fc196688 Local static initializers are evil, avoid them like plague.
git-svn-id: http://pcsx2.googlecode.com/svn/trunk@4304 96395faa-99c1-11dd-bbfe-3dabce05a288
2011-02-16 03:19:36 +00:00
gabest11 1e6f280021 - more project cleanups and small code changes, also added the psx emu interface again
- someone should check __xgetbv under linux (avx/fma detection)

git-svn-id: http://pcsx2.googlecode.com/svn/trunk@4295 96395faa-99c1-11dd-bbfe-3dabce05a288
2011-02-12 21:45:16 +00:00
sudonim1 4457fe40fc Removed all trailing whitespace in *.c *.cpp *.h because it irritates me.
git-svn-id: http://pcsx2.googlecode.com/svn/trunk@2897 96395faa-99c1-11dd-bbfe-3dabce05a288
2010-04-25 00:31:27 +00:00
gabest11 83b15dbac0 GSdx:
- trying the dx10.1-only "gather" shader instruction for palletized lookups ("8-bit texture" mode), saves 4 instructions which isn't much but still... (not tested, don't have ati)
- may fix the intel gma "no output" bug (don't have gma either :P)
- and the usual small code optimizations


git-svn-id: http://pcsx2.googlecode.com/svn/trunk@1549 96395faa-99c1-11dd-bbfe-3dabce05a288
2009-07-22 03:55:28 +00:00
gabest11 f9f056d581 GSdx: fixing/breaking things again... palletized texture lookup can be done by pixel shader now (selectable, off by default), if you have a fast card it may help with texture heavy games, otherwise it is only going to be slower.
git-svn-id: http://pcsx2.googlecode.com/svn/trunk@1472 96395faa-99c1-11dd-bbfe-3dabce05a288
2009-07-06 16:35:06 +00:00
gabest11 4c76909afe GSdx: small optimizations and tried to fix that dx9 fullscreen alt+tab crash
git-svn-id: http://pcsx2.googlecode.com/svn/trunk@1463 96395faa-99c1-11dd-bbfe-3dabce05a288
2009-07-04 15:14:04 +00:00
gabest11 5f45219eb9 GSdx: fixes for ffx2 menu
git-svn-id: http://pcsx2.googlecode.com/svn/trunk@1412 96395faa-99c1-11dd-bbfe-3dabce05a288
2009-06-20 20:28:36 +00:00
gabest11 5736bbd449 GSdx: looking for an opengl guru...
git-svn-id: http://pcsx2.googlecode.com/svn/trunk@1360 96395faa-99c1-11dd-bbfe-3dabce05a288
2009-06-12 19:09:17 +00:00
gabest11 5648d07b96 GSdx: more refactoring (aka search and replace)
git-svn-id: http://pcsx2.googlecode.com/svn/trunk@1179 96395faa-99c1-11dd-bbfe-3dabce05a288
2009-05-14 16:41:52 +00:00
gabest11 c64691811a GSdx: 3.4.9. Texture Function => Decal/TCC=RGB: Av = At (gs_user v3 and v5) or Av = Af (v6)? Seems it was Af... Sonic unleashed was one of the games relying on this setting, could not find any other yet. Also removed the green stripes, shadow sw-mode only.
git-svn-id: http://pcsx2.googlecode.com/svn/trunk@1145 96395faa-99c1-11dd-bbfe-3dabce05a288
2009-05-07 12:34:58 +00:00
gabest11 120971ec4f GSdx: Implemented edge anti-aliasing (aa1) for software mode, bios or ffx are good test subjects (not many other games use it). It's still a bit slow but could be improved a lot by not doing 4 pixels with sse for each single edge pixel, that's just a lot of unnecessary texture lookups. The bios config screen cubes are still bogus, gs_user on aa1 isn't too helpful...
git-svn-id: http://pcsx2.googlecode.com/svn/trunk@721 96395faa-99c1-11dd-bbfe-3dabce05a288
2009-03-09 01:42:56 +00:00
gabest11 e934016121 GSdx: vtune JIT code profiling support (somebody should implement this in pcsx2 too, see iJIT_NotifyEvent(iJVM_EVENT_TYPE_METHOD_LOAD_FINISHED, ...), plzkthx :P) and other minor fixes/optimizations.
git-svn-id: http://pcsx2.googlecode.com/svn/trunk@570 96395faa-99c1-11dd-bbfe-3dabce05a288
2009-02-22 16:28:39 +00:00
gabest11 b3cb4ad3f8 GSdx: completed the ps1 renderer conversion for xbyak, GSRasterizer::Draw* functions should be next.
git-svn-id: http://pcsx2.googlecode.com/svn/trunk@550 96395faa-99c1-11dd-bbfe-3dabce05a288
2009-02-20 10:59:10 +00:00