preserve-emulation/ppsspp
1
0
Fork 0
mirror of https://github.com/hrydgard/ppsspp.git synced 2026-07-24 15:54:45 +02:00
Code Issues Packages Projects Releases Wiki Activity
33,136 Commits 77 Branches 131 Tags
fb875b94de26c9900d38f412ce96af9860a248b3
Commit Graph
7472 Commits
Author SHA1 Message Date
Lubos 5be6665df8 Merge branch 'master' into feature_openxr_stereo 2022-09-04 12:37:41 +02:00
Lubos 1df1d79b9e OpenXR - Stereo without multiview added 2022-09-04 12:36:52 +02:00
Henrik Rydgård 3d2e78fbf4 spaces->tabs 2022-09-04 00:11:22 +02:00
Henrik Rydgård ec173559f8 Additional cleanup 2022-09-04 00:10:55 +02:00
Henrik Rydgård ceda7aef49 Add log functions to state structs 2022-09-04 00:10:55 +02:00
Henrik Rydgård d35749bfc8 Move the computation of simulate-logic shader type earlier 2022-09-04 00:10:55 +02:00
Henrik Rydgård 87b14fe1c2 Some more renaming and refactoring, extracted from a future PR 2022-09-04 00:10:55 +02:00
Henrik Rydgård 8e39018b95 Channel mask refactor 2022-09-04 00:10:54 +02:00
Henrik Rydgård 5f1ffeae70 Correct LogicOp API support detection 2022-09-04 00:08:40 +02:00
Henrik Rydgård d3309dd8e9 Minor refactor with pipelines in QueueRunner (makes more information available for easier debugging) 2022-09-03 23:08:02 +02:00
Henrik Rydgård 6b1e4806cf Don't parallelize shader compiles in debug mode, sync in debug allocator kills performance. 2022-09-03 23:07:39 +02:00
Henrik RydgårdandGitHub 14465cee13 Merge pull request #15951 from unknownbrackets/softgpu-rectfast 
Improve rectangle fast path checks in softgpu
 2022-09-03 22:58:34 +02:00
Unknown W. Brackets 15b73b56bb softgpu: Correct texture check in fastpath. 
This is the no textures option, so we shouldn't check sampler state at
all.  Presumably this was failing and reading uninitialized data...
 2022-09-03 11:21:14 -07:00
Unknown W. Brackets af004dd53d softgpu: Validate alpha test properly for fastpath. 
We don't want a weird LESS test being mishandled.
 2022-09-03 11:19:24 -07:00
Unknown W. Brackets ba1ced4992 softgpu: Avoid rect fast path with subpixel offset. 
It doesn't handle it right, see #15876.  We still handle these with
DrawRectangle(), which gets it right.
 2022-09-03 11:08:11 -07:00
Henrik RydgårdandGitHub b92ea74fb3 Merge pull request #15589 from hrydgard/glsl-compilation-parallelization 
Vulkan: Parallelize GLSL compilation
 2022-09-03 17:00:01 +02:00
Henrik Rydgård fb3f417e77 Fix lying comment, log level reduction 2022-09-03 16:15:46 +02:00
Henrik Rydgård c06cf8efaa Switch Vulkan pipelines to use promises for synchronization 
Slightly more expensive I guess but shouldn't be much of a bottleneck.
 2022-09-03 15:21:08 +02:00
Henrik Rydgård 6e605fa2bc Fix silly logic, fixes the shader leak. 2022-09-03 15:21:07 +02:00
Henrik Rydgård 8775837c11 Work towards finding the leak 2022-09-03 15:21:07 +02:00
Henrik Rydgård 93422f6dea Don't block on shader creation until it's time to create the pipeline. 2022-09-03 15:21:07 +02:00
Henrik Rydgård e2c740827a Use shader module promises as keys in pipeline cache instead of the actual shader modules. 
Prevents early block.
 2022-09-03 15:21:07 +02:00
Henrik Rydgård fc418ee16f Create shader modules (and compile GLSL) on worker threads 2022-09-03 15:21:07 +02:00
Henrik Rydgård ec0586ba50 Use promises to hold shader modules, for later asyncification 2022-09-03 15:21:07 +02:00
Henrik Rydgård c846c2dfa8 Remove confusing resetFramebufferRead flag from secondary framebuffer binding 2022-09-03 14:48:07 +02:00
Henrik Rydgård 8c56abd3dc Remove setting "Disable slower framebuffer effects" 2022-09-03 11:06:43 +02:00
Henrik Rydgård 68d88726ae Remove unused function 2022-09-03 11:06:43 +02:00
Henrik RydgårdandGitHub 88694bde0c Merge pull request #15943 from unknownbrackets/softgpu-vertex 
vertexjit: Clamp through float pos during decode
 2022-09-03 07:02:46 +02:00
Henrik Rydgård fd6ab4c495 Avoid recomputing the shaderblend setup in ComputeFragmentShaderID. 2022-09-02 23:06:41 +02:00
Henrik Rydgård 0fe064d970 Small steps 2022-09-02 22:43:58 +02:00
Henrik Rydgård f81a28f661 Some reordering 2022-09-02 22:40:15 +02:00
Henrik Rydgård de712b7804 More renaming 2022-09-02 22:20:11 +02:00
Henrik Rydgård b84cda2876 Plumb the computed pipeline state into ComputeFragmentShaderID. 2022-09-02 22:16:57 +02:00
Henrik Rydgård c784c0e94b Some renaming. Store the BlendState/MaskState. 2022-09-02 21:07:29 +02:00
Unknown W. Brackets 470d2f0f4e vertexjit: Remove unused ReadUV() cases. 2022-09-01 23:33:52 -07:00
Unknown W. Brackets f2d5d668a5 vertexjit: Remove unused ReadPos() cases. 2022-09-01 23:33:52 -07:00
Unknown W. Brackets 9f48ddb18c vertexjit: Clamp through float pos during decode. 
Rather than decoding it on read, better to decode it right away.
 2022-09-01 23:33:52 -07:00
Unknown W. Brackets c5fbb400de Replacement: Detect OOM even with replacements. 
Also allow saving video textures if so configured.
 2022-09-01 19:05:20 -07:00
Unknown W. Brackets 1033fa6cd7 Vulkan: Prevent saving tex during replace load. 2022-09-01 18:52:52 -07:00
Lubos 00aaf9e1a6 Merge branch 'master' into feature_openxr_stereo 2022-09-01 18:09:21 +02:00
Henrik Rydgård 2bf0c86ac7 Fix image format when dumping textures. 
Fixes #15831
 2022-09-01 15:04:59 +02:00
Henrik Rydgård 11d4f9acb6 Fix texture saving for Vulkan 2022-09-01 15:04:59 +02:00
Henrik Rydgård 9097fdaae6 Shrink VkRenderData from 88 to 64 bytes. 2022-09-01 14:21:34 +02:00
Henrik Rydgård ecb84987ef Remove allocation from the heaviest MemBlockInfo path (BlockTransfer). Required some refactoring. 2022-09-01 11:59:35 +02:00
Henrik Rydgård c9048c3748 Shrink the DeferredDrawcall struct, because why not. Assorted cleanup 2022-09-01 11:59:33 +02:00
Henrik Rydgård a854fbbe76 fbtex matching: Shrink the matching struct, use TinySet instead of vector to avoid an allocation. 2022-09-01 10:24:52 +02:00
Henrik RydgårdandGitHub 608710f9f5 Merge pull request #15936 from hrydgard/framebuffer-cleanups 
Framebuffer settings and block transfer speedup
 2022-09-01 10:04:31 +02:00
Henrik Rydgård 1c9ec36fd8 Stricter checks for negative framebuffer offsets. Fixes #15937 2022-09-01 08:26:23 +02:00
Henrik Rydgård 470efac7d8 Merge GetFramebufferCandidates and GetBestCandidateIndex into one function. 2022-09-01 00:35:43 +02:00
Henrik Rydgård a8a9fb4206 Much faster framebuffer matching (no more ToString every time there's more than 1 candidate, etc). 2022-08-31 14:23:56 +02:00