137 Commits

Author SHA1 Message Date
Unknown W. Brackets
b33d0c4046 softgpu: Use SSE for texture sampling. 2014-03-16 14:33:42 -07:00
Unknown W. Brackets
b357b00ace softgpu: Use SSE for through texture coords. 2014-03-16 14:30:20 -07:00
Unknown W. Brackets
dd140b73bb softgpu: Use SSE for gouraud shading. 2014-03-16 14:29:22 -07:00
chinhodado
d110d4e3cf change to correct enums 2014-02-20 21:29:00 -05:00
Unknown W. Brackets
5a7d400f4b softgpu: Fix lines drawn upward or leftward. 2014-01-19 23:32:53 -08:00
Unknown W. Brackets
8271ec1a55 softgpu: Avoid a divide by 0 in line interpolation.
Fixes crash in Persona 2 loading screens.
2014-01-19 23:32:12 -08:00
Unknown W. Brackets
74c1b21b59 softgpu: Don't redraw every pixel of a line 16x.
Fixes the insane slowness when processing lines.  Now it's just regular
super slow.
2014-01-19 23:28:34 -08:00
Unknown W. Brackets
4bb6a77519 softgpu: Cleanup DrawPoint a bit.
Mostly in DrawSinglePixel() now.
2014-01-18 20:50:38 -08:00
Unknown W. Brackets
8f5fc4f079 softgpu: Interpolate/texture lines and points.
Also, expand out pixels in lines, it's just too slow otherwise (30%
overhead or so.)
2014-01-18 20:48:48 -08:00
Unknown W. Brackets
e216032a8c softgpu: Separate out texturing logic.
This way it can be applied to other primitives.  But, this will probably
need larger changes to implement mipmapping.
2014-01-18 20:10:42 -08:00
raven02
4c25bb0365 softgpu: Let's try doubling before alpha blend.
Fixes #5122.
2014-01-18 19:58:28 -08:00
Unknown W. Brackets
08eecba6ba softgpu: Refactor pixel drawing to avoid code dup. 2014-01-18 19:57:12 -08:00
Unknown W. Brackets
01090f4ce7 softgpu: Respect the texture filering option.
Nearest is much faster, so you can force it to nearest for a good boost in
speed.
2014-01-18 16:39:46 -08:00
Unknown W. Brackets
b9f4cc4481 Double colors only when texturing, and after test.
Per tests.
2014-01-12 11:45:29 -08:00
Unknown W. Brackets
eb0ea71f24 softgpu: Use region for debugger buffer previews.
It's fine if it resizes or etc., but this should generally be the
interesting area of vram to show, without any confusing noise on the
side/etc.
2013-12-29 13:02:04 -08:00
Unknown W. Brackets
9f86959826 Fix typesafety (ScreenCoords not DrawingCoords.) 2013-12-29 12:59:26 -08:00
Unknown W. Brackets
9aa1b15569 softgpu: Fix black objects with bilinear filtering.
Dividing by 256 gives us a range of 0-253, and incorrectly gets zero for
some components (guessing alpha, primarily?) in some games.

Dividing by 255 has a speed hit, so a slight bias to the top left seems
like a decent performance compromise (and it's a very slight bias.)
2013-12-15 11:59:25 -08:00
Unknown W. Brackets
96d6f52a13 softgpu: Retain accuracy in depth for tests.
Fixes intro/menu issues in Phantasy Star Portable 1/2, and text in the
first, both of which now look correct.
2013-12-15 11:59:25 -08:00
Unknown W. Brackets
6ca5662b03 softgpu: Properly swizzle 16 bit textures.
Fixes for example the graphics in Fieldrunners.
2013-12-15 11:59:24 -08:00
Unknown W. Brackets
2f0c8c2877 softgpu: Attempt to implement GE_PROJMAP_UV.
Looks okay, not sure if it's fully correct.
2013-12-15 11:59:22 -08:00
Henrik Rydgard
d4eee74b54 SoftGPU: Prepare for mipmapping. Getting the uv deltas seems tricky though... 2013-12-10 15:33:59 +01:00
raven02
2e7609b084 Cleanup ClearmodeDepthMask 2013-12-10 20:48:20 +08:00
Henrik Rydgård
20873f2985 Remove unnecessary ifdef 2013-12-09 16:53:46 +01:00
raven02
75492719e8 std::bind build fix for VS2012 2013-12-09 21:06:38 +08:00
Henrik Rydgård
68b08ffb9f SoftGPU: Quick and dirty lines and points implementation
No line/point texturing support yet.
2013-12-09 12:43:49 +01:00
Unknown W. Brackets
c6a441965e Move shared clut check to gstate. 2013-12-08 23:11:56 -08:00
Unknown W. Brackets
cdccf06b62 softgpu: Avoid some ScreenToDrawing() overhead. 2013-12-08 22:38:09 -08:00
Unknown W. Brackets
30394df775 softgpu: Make texture lookups do less math. 2013-12-08 22:37:45 -08:00
Henrik Rydgard
bd1d4e2d92 As if softgpu wasn't slow enough already, implement bilinear filtering. 2013-12-09 01:44:23 +01:00
Unknown W. Brackets
637a9de75d Optimize out some clear mode checks.
Since it's really common and the rasterizer is slow for large draws.
2013-12-08 14:24:17 -08:00
Unknown W. Brackets
598d509fb9 Use multiple threads if available for softgpu.
VIVA OVERHEAD.  But it's actually generally faster this way, even with
only 2 cores running (but not a huge improvement then.)

On an i7, this improves things ~2.5x with 8 cores.
2013-12-08 13:41:20 -08:00
Henrik Rydgård
428b27903b Merge pull request #4556 from unknownbrackets/ctrl-timing
Adjust ctrl timing (sample and wake after vblank)
2013-11-18 05:06:58 -08:00
Unknown W. Brackets
7a9fe955f8 softgpu: Only check clear mode once for depthtest.
Small, small optimization.  It's only called outside clear mode.
2013-11-17 19:32:21 -08:00
Unknown W. Brackets
dfadb67ea1 Avoid some operator overloads.
Causing ambiguity.
2013-11-17 14:42:58 -08:00
Unknown W. Brackets
1311d71455 softgpu: Fix reversed stencil test funcs.
LESS applies to the ref compared to the stencil, not the stencil compared
to the ref.

With this, stencil testing is finally being done properly in Star Ocean.
2013-11-10 03:21:01 -08:00
Unknown W. Brackets
2db98b8669 softgpu: Don't apply logic op to stencil value.
Also, they were not at all applying correctly.
2013-11-10 03:21:00 -08:00
Unknown W. Brackets
438928de40 softgpu: Apply stencil op even when depth disabled. 2013-11-10 03:21:00 -08:00
Unknown W. Brackets
5327ce80d5 softgpu: throughmode doesn't disable any tests. 2013-11-10 03:20:59 -08:00
Unknown W. Brackets
5751e0a0a4 softgpu: Simplify color clamping. 2013-11-10 03:20:58 -08:00
Unknown W. Brackets
21deaef35a Invert gstate.getClearModeColorMask().
It makes more sense matching gstate.getColorMask().
2013-11-10 03:20:58 -08:00
Unknown W. Brackets
5817684e28 softgpu: Never write alpha, only stencil. 2013-11-10 03:20:57 -08:00
Unknown W. Brackets
be9fe9ad9a softgpu: Fix stencil read/write in some formats. 2013-11-10 03:20:56 -08:00
Unknown W. Brackets
c1897b9e54 Compare stencil correctly based on bitdepth.
All sorts of funny rules.  See comments here:
https://github.com/hrydgard/ppsspp/pull/4150
2013-10-12 22:20:38 -07:00
Unknown W. Brackets
b9811c7c97 Use the slow way to get the softgpu texture.
So that it actually shows what the softgpu is using.
2013-10-07 01:26:48 -07:00
Unknown W. Brackets
5c2d418d95 Read softgpu stencil using existing funcs.
Slower but at least it's the same.
2013-10-07 01:02:58 -07:00
Unknown W. Brackets
b9453f82ca Use the full alpha bits for stencil in softgpu.
It works to increment and set the stencil, so it can't be high bits only.
2013-10-07 01:00:43 -07:00
Unknown W. Brackets
013e3f3be2 Add very basic and slow DXT decoding to softgpu. 2013-10-05 12:25:18 -07:00
Unknown W. Brackets
e79781b1a3 softgpu: Skip the stencil/depth tests in a clear. 2013-09-28 12:15:07 -07:00
Henrik Rydgard
82a2f6443d Oops (last commit) 2013-09-21 21:05:15 +02:00
Henrik Rydgard
34341b0373 Minor optimization in swrast 2013-09-21 21:03:43 +02:00