preserve-emulation/ppsspp
1
0 Fork 0
mirror of https://github.com/hrydgard/ppsspp.git synced 2026-11-14 18:19:36 +01:00
Code Issues Packages Projects Releases Wiki Activity
43,966 Commits 85 Branches 131 Tags
http-split-headers
Commit Graph
9475 Commits
Author SHA1 Message Date
Henrik Rydgård d9f74d2fb7 ivec->uvec, comment fix 2022-09-26 13:05:25 +02:00
Henrik Rydgård fc30b04430 ShaderUniforms: cleanup, put every "4-float" on a line for clarity 2022-09-26 13:05:25 +02:00
Henrik Rydgård cfa427c37a Shuffle constants around, squeezing them into gaps. Saves another 16 bytes. 2022-09-26 13:05:24 +02:00
Henrik Rydgård f4b71e2dc7 Fragment shader uniforms: Pack color mask in 32 bits instead of expand to 128 bits. 
Allows us to save 16 bytes from the main uniform buffer, since there's
free 32-bit spaces here and there to use.
 2022-09-26 13:04:56 +02:00
Henrik Rydgård 07ca9e4656 Fold the "materialUpdate" flag into the light ubershader part. 
This reduces the number of vertex shaders and thus pipelines by quite a
bit more in a few games, like Tekken and GoW, continuing the fight
against shader compile stutter.

The perf impact should be minimal if not positive due to less pipeline
changes.

GLES fixes

Make the vertex input declarations match (always declare fog input).  Fixes D3D11 validation

Tess fix
 2022-09-26 12:06:16 +02:00
Henrik Rydgård ad1021ea4b Add some recent flags to FragmentShaderDesc 2022-09-26 12:06:16 +02:00
Henrik Rydgård 76f03d30bf Remove suspicious dirty flag 2022-09-26 11:21:40 +02:00
Henrik Rydgård 196f8e3461 Prepare for dynamic mat update 2022-09-26 11:21:40 +02:00
Henrik Rydgård 94e439280e Merge pull request #16111 from hrydgard/always-compute-fog-in-vs 
Always do the vertex shader part of the fog computation.
 2022-09-26 11:20:06 +02:00
Henrik Rydgård 9d1355e137 Always do the vertex shader part of the fog computation. 
In #16104, we drastically reduced the number of shader variants for
games that use flexible lighting setups. I looked at a few games and it
seems that a lot of games have the same shaders with fog on/off, while
fog is super cheap to compute. So let's just always do it, reducing
vertex shader variants further (though the amount of pipelines will probably
remain the same, since we still specialize the fragment shader).

Might also be worth adding a dynamic bool for the fragment shader, but
if so, doing it separately.
 2022-09-26 09:30:54 +02:00
Unknown W. Brackets c80f325912 GPU: Fix SSE4 Vec3f normalize. 
Was sometimes adding in garbage data, which could create NANs.
 2022-09-26 00:24:12 -07:00
Henrik Rydgård f30252f8d5 Oops! Also, testfix 2022-09-25 23:35:08 +02:00
Henrik Rydgård 96f054f098 Fix light ubershader for D3D11 and OpenGL, GLES unsigned/signed stuff 2022-09-25 23:35:08 +02:00
Henrik Rydgård 7adba20fac Experiment: Generate "Ubershaders" that can handle all lighting configurations 
This drastically reduces the shader compile stutter that happens when a lot of new
light setups are created, like on the first punch in Tekken 6.

There's more stuff that might benefit from being made dynamic like this.
These branches are very cheap on modern GPUs since they're branching on
a uniform variable, so no divergence.

Only tested on Vulkan. I think we'll need to keep the old path too for
gpus like Mali-450...
 2022-09-25 23:35:01 +02:00
Henrik Rydgård b1afeeaf43 Merge pull request #16100 from unknownbrackets/d3d9-debugger 
D3D9: Allow INTZ depth buffers more correctly
 2022-09-25 17:37:56 +02:00
Unknown W. Brackets e6db0bef2d Merge pull request #16099 from hrydgard/vulkan-dont-always-alloc-depth 
Vulkan: Avoid allocating depth images for stuff like temp copies, depal buffers etc.
 2022-09-25 08:05:50 -07:00
Henrik Rydgård a26a353c25 Merge pull request #16102 from unknownbrackets/softgpu-bin-tweaks 
softgpu: Avoid waiting for a thread to drain
 2022-09-25 10:01:44 +02:00
Henrik Rydgård 70c5ca62e6 Remove debug log. Add some new debug log though, unrelated to this PR, for fb clut + fb texture. Plus a couple asserts. 2022-09-25 09:56:39 +02:00
Henrik Rydgård 8a4147c042 Merge pull request #16101 from unknownbrackets/softgpu-fixes 
softgpu: Avoid fast path in another wrong case
 2022-09-25 09:52:29 +02:00
Unknown W. Brackets 24560eef5c softgpu: Avoid waiting for a thread to drain. 
If we can, we want to keep the thread queues healthy, but not full.
Reduce the amount we push on a typical drain to avoid the Wait().
 2022-09-24 20:01:00 -07:00
Unknown W. Brackets 1aa6841759 softgpu: Increase queued prims. 
We made them smaller, so we can queue more of them in the same space.
Helps a little bit.
 2022-09-24 20:01:00 -07:00
Unknown W. Brackets 444781c7b0 softgpu: Fix triangle strip with partial rects. 
Seen in Wild Arms XF shop menu.
 2022-09-24 18:55:45 -07:00
Unknown W. Brackets c47d7eab38 softgpu: Simply 5551 blending fast path. 
Since it only supports multiply and add, let's just stick with that.
 2022-09-24 18:55:45 -07:00
Unknown W. Brackets 1eeb4f0bcf softpu: Refactor out 5551 fast path checks. 
They were duplicated, and better to organize them according to state.
 2022-09-24 18:55:45 -07:00
Unknown W. Brackets f30b1d048d softgpu: Avoid fast path in another wrong case. 
Seen in Kurohyo.  Missed that the alpha blend check essentially means only
standard blending can work.
 2022-09-24 17:53:09 -07:00
Unknown W. Brackets 81e8336985 D3D9: Allow INTZ depth buffers more correctly. 
The FBO check was wrong and just always failed.
 2022-09-24 15:17:18 -07:00
Henrik Rydgård 08d2cb4486 Bump the shader cache version 2022-09-24 22:40:42 +02:00
Henrik Rydgård 9f3dfe7ebe Vulkan: Don't compile pipeline variants that don't make sense given their flags. 
Ran into this with cache files from previous version of my change.

Also bumping the shader cache ID again to avoid this in other ways, but
good to be robust here.
 2022-09-24 22:39:22 +02:00
Henrik Rydgård c3b4caa30b Merge pull request #15984 from lvonasek/compat_openxr_gta 
OpenXR - Sky fix for GTA games
 2022-09-24 17:16:28 +02:00
Unknown W. Brackets c76d31dfa8 GPU: Cleanup unused CheckAlpha() funcs. 2022-09-24 02:00:03 -07:00
Unknown W. Brackets 6e6535c263 softjit: Skip reading dst pixel where blended out. 
Sometimes used by blends used purely to multiply the source color by
something, usually prep for bloom.
 2022-09-24 02:00:03 -07:00
Unknown W. Brackets a4c3718431 softgpu: Optimize rectangle sampling/blending. 
Sometimes the vertex color or alpha can allow us to optimize away some
multiplication.
 2022-09-24 02:00:03 -07:00
Unknown W. Brackets 794a5c07ad softgpu: Ignore a needless color test case. 
This happens in Ridge Racer, and we can entirely skip the color test.
 2022-09-24 02:00:03 -07:00
Unknown W. Brackets 7ff5434968 GE Debugger: Tag frame dump replay VRAM writes. 
Just for debugging, it's helpful especially paired with softgpu tagging.
 2022-09-23 21:20:14 -07:00
Unknown W. Brackets c3c5450b8f GE Debugger: Fix small tex/clut recopying. 
If it's less than 256 bytes, we can't mark the entire VRAM area copied.
This still helps frame dumps avoid excessively slow VRAM recopying
situations, but fixes issues like missing trees in #12738.
 2022-09-23 21:18:39 -07:00
Unknown W. Brackets b56bd0d0fc Merge pull request #16090 from hrydgard/more-vulkan-cleanup-work 
Simplify synchronization in VulkanRenderManager
 2022-09-23 17:24:34 -07:00
Henrik Rydgård d743bfac93 Merge pull request #16085 from unknownbrackets/softgpu-vert 
softgpu: Cache reused indexed verts
 2022-09-24 00:00:26 +02:00
Henrik Rydgård 1259283c2e More tweaks, fix crash on exit (double-join thread) 2022-09-23 22:10:29 +02:00
Lubos adffbb2ea7 Merge branch 'master' into compat_openxr_gta 2022-09-23 14:16:58 +02:00
Henrik Rydgård 7884e4ccb3 Another uninitialized variable (VAI minihash/hash) 2022-09-23 12:33:16 +02:00
Henrik Rydgård ac7ca963db Make valgrind happy 2022-09-23 12:24:43 +02:00
Henrik Rydgård bb6919ebcb Merge pull request #16087 from unknownbrackets/depth-upload 
GPU: Upload depth only on first usage
 2022-09-23 09:07:33 +02:00
Unknown W. Brackets 93c909a88e GPU: Upload depth only on first usage. 
Fixes various glitches in Kingdom Hearts, etc.
 2022-09-23 00:04:14 -07:00
Unknown W. Brackets 66b6dfd0a5 softgpu: Fix self-render detect in Ridge Racer. 
When we flush we mark all pending writes zero, but we rely on this being
set to detect self-render.

TRANSFORM_ALL was wrong as well, sometimes clearing BINNER_RANGE.
 2022-09-22 20:36:15 -07:00
Unknown W. Brackets 88b3b26ed3 softgpu: Cache reused indexed verts. 
This happens a lot for spline/bezier, so can significantly speed up curve
heavy scenes.  Isn't necessarily that common otherwise, though.
 2022-09-22 18:27:59 -07:00
Unknown W. Brackets 067fac6817 softgpu: Skip matrix multiply for fog factor calc. 
We can just use a dot product instead, and always skip viewpos.
 2022-09-22 18:19:53 -07:00
Unknown W. Brackets 84a3f6de71 softgpu: Remove unnecessary state param. 
Oops, meant to remove this when refactoring imm prims.
 2022-09-22 18:18:49 -07:00
Henrik Rydgård a6d6e0a3cc Texture/Framebuffer match: Ignore stride if texHeight == 1. Fixes Ridge Racer lens flares. 2022-09-22 22:11:16 +02:00
Henrik Rydgård 078fa9beb2 Fix corruption of Ridge Racer speedometers with AutoMaxQuality enabled. 
See #8509
 2022-09-22 15:27:17 +02:00
Henrik Rydgård c3cbb68452 Merge pull request #16072 from hrydgard/depth-free-renderpass 
Vulkan: Don't have renderpasses store/load depth buffers when we don't use them
 2022-09-22 11:05:25 +02:00