Commit graph

1947 commits

Author SHA1 Message Date
bunnei
e1ad08d11b Merge pull request #3098 from ReinUsesLisp/shader-invalidations
gl_shader_cache: Miscellaneous changes to shaders
2019-11-24 19:36:30 -05:00
bunnei
87748a8965 Merge pull request #3141 from ReinUsesLisp/gl-position
gl_shader_gen: Apply default value to gl_Position
2019-11-23 13:23:46 -05:00
ReinUsesLisp
d91f11ca92 gl_device: Reserve base bindings on limited devices
SSBOs and other resources are limited per pipeline on Intel and AMD.
Heuristically reserve resources per stage having in mind the reported
OpenGL limits.
2019-11-22 21:28:50 -03:00
ReinUsesLisp
fb6a01233f gl_state: Skip null texture binds
glBindTextureUnit doesn't support null textures. Skip binding these.
2019-11-22 21:28:50 -03:00
ReinUsesLisp
d892e6789a gl_rasterizer: Disable compute shaders on Intel
Intel's proprietary driver enters in a corrupt state when compute
shaders are executed. For now, disable these.
2019-11-22 21:28:50 -03:00
ReinUsesLisp
2726dca44c gl_shader_cache: Hack shared memory size
The current shared memory size seems to be smaller than what the game
actually uses. This makes Nvidia's driver consistently blow up; in the
case of FE3H it made it explode on Qt's SwapBuffers while SDL2 worked
just fine. For now keep this hack since it's still progress over the
previous hardcoded shared memory size.
2019-11-22 21:28:49 -03:00
ReinUsesLisp
a52b0f1874 gl_shader_decompiler: Normalize image bindings 2019-11-22 21:28:49 -03:00
ReinUsesLisp
99ed7e5128 gl_shader_decompiler: Normalize cbuf bindings
Stage and compute shaders were using a different binding counter.
Normalize these.
2019-11-22 21:28:49 -03:00
ReinUsesLisp
54b16fb27a gl_rasterizer: Add missing cbuf counter reset on compute 2019-11-22 21:28:49 -03:00
ReinUsesLisp
0ecd1a227d gl_shader_cache: Remove dynamic BaseBinding specialization 2019-11-22 21:28:49 -03:00
ReinUsesLisp
5f6a589d63 video_core: Unify ProgramType and ShaderStage into ShaderType 2019-11-22 21:28:48 -03:00
ReinUsesLisp
2930400adc gl_rasterizer: Bind graphics images to draw commands
Images were not being bound to draw invocations because these would
require a cache invalidation.
2019-11-22 21:28:48 -03:00
ReinUsesLisp
46ad4bd297 gl_shader_cache: Specialize local memory size for compute shaders
Local memory size in compute shaders was stubbed with an arbitary size.
This commit specializes local memory size from guest GPU parameters.
2019-11-22 21:28:48 -03:00
ReinUsesLisp
1669dfa793 gl_shader_cache: Specialize shared memory size
Shared memory was being declared with an undefined size. Specialize from
guest GPU parameters the compute shader's shared memory size.
2019-11-22 21:28:47 -03:00
ReinUsesLisp
b0b505984d gl_shader_cache: Specialize shader workgroup
Drop the usage of ARB_compute_variable_group_size and specialize compute
shaders instead. This permits compute to run on AMD and Intel
proprietary drivers.
2019-11-22 21:28:47 -03:00
ReinUsesLisp
77f86f48ac shader/texture: Deduce texture buffers from locker
Instead of specializing shaders to separate texture buffers from 1D
textures, use the locker to deduce them while they are being decoded.
2019-11-22 21:28:47 -03:00
ReinUsesLisp
68cdb253db gl_shader_gen: Apply default value to gl_Position
Nvidia has sane default output values for varyings, but the other
vendors don't apply these. To properly emulate this we would have to
analyze the shader header. For the time being, apply the same default
Nvidia applies so we get the same behaviour on non-Nvidia drivers.
2019-11-19 20:32:01 -03:00
bunnei
44d462d511 Merge pull request #3086 from ReinUsesLisp/format-lookups
texture_cache: Use a flat table instead of switch for texture format lookups
2019-11-19 18:29:17 -05:00
bunnei
3546633a63 Merge pull request #3047 from ReinUsesLisp/clip-control
gl_rasterizer: Emulate viewport flipping with ARB_clip_control
2019-11-15 12:09:19 -05:00
ReinUsesLisp
c6afe25477 texture_cache: Drop abstracted ComponentType
Abstracted ComponentType was not being used in a meaningful way.
This commit drops its usage.

There is one place where it was being used to test compatibility between
two cached surfaces, but this one is implied in the pixel format.
Removing the component type test doesn't change the behaviour.
2019-11-14 18:21:42 -03:00
Fernando Sahmkow
dfaeb0a97d Merge pull request #3081 from ReinUsesLisp/fswzadd-shuffles
shader: Implement FSWZADD and reimplement SHFL
2019-11-14 10:27:27 -04:00
Rodrigo Locatti
e4a24b0c80 Merge pull request #3084 from ReinUsesLisp/cast-warnings
video_core: Treat implicit conversions as errors
2019-11-13 02:16:22 -03:00
bunnei
bb464b6df2 Merge pull request #3082 from ReinUsesLisp/fix-lockers
gl_shader_cache: Fix locker constructors
2019-11-09 13:58:36 -05:00
ReinUsesLisp
905cc250a4 video_core: Silence implicit conversion warnings 2019-11-08 22:48:50 +00:00
bunnei
4eb49118d9 Merge pull request #3080 from FernandoS27/glsl-fix
GLSLDecompiler: Correct Texture Gather Offset.
2019-11-08 15:56:29 -05:00
ReinUsesLisp
cee3db83d7 gl_shader_cache: Fix locker constructors
Properly pass engine when a shader is being constructed from memory.
2019-11-07 20:43:31 -03:00
ReinUsesLisp
c8988f3e46 gl_shader_cache: Enable extensions only when available
Silence GLSL compilation warnings.
2019-11-07 20:08:42 -03:00
ReinUsesLisp
1b2de9b6b6 gl_shader_decompiler: Add safe fallbacks when ARB_shader_ballot is not available 2019-11-07 20:08:42 -03:00
ReinUsesLisp
bb94bcc991 shader_ir/warp: Implement FSWZADD 2019-11-07 20:08:41 -03:00
ReinUsesLisp
5fc04875a1 gl_shader_decompiler: Reimplement shuffles with platform agnostic intrinsics 2019-11-07 20:08:41 -03:00
Fernando Sahmkow
a4d70a6eac GLSLDecompiler: Correct Texture Gather Offset.
This commit corrects the argument ordering in textureGatherOffset.
2019-11-07 11:43:56 -04:00
ReinUsesLisp
07fa50c92a gl_rasterizer: Remove front facing hack 2019-11-07 01:52:18 -03:00
ReinUsesLisp
56604999fe gl_shader_decompiler: Fix typo "y_negate"->"y_direction" 2019-11-07 01:52:18 -03:00
ReinUsesLisp
02493e1b4d gl_shader_manager: Remove unused variable in SetFromRegs 2019-11-07 01:52:18 -03:00
ReinUsesLisp
1bb64655cb gl_rasterizer: Emulate viewport flipping with ARB_clip_control
Emulates negative y viewports with ARB_clip_control. This allows us to
more easily emulated pipelines with tessellation and/or geometry shader
stages. It also avoids corrupting games with transform feedbacks and
negative viewports (gl_Position.y was being modified).
2019-11-07 01:52:18 -03:00
bunnei
e2df5620dd Merge pull request #3057 from ReinUsesLisp/buffer-sub-data
gl_rasterizer: Upload constant buffers with glNamedBufferSubData
2019-11-06 10:08:55 -05:00
Rodrigo Locatti
afa4528b51 Merge pull request #3039 from ReinUsesLisp/cleanup-samplers
shader/node: Unpack bindless texture encoding
2019-11-06 04:54:11 +00:00
ReinUsesLisp
623b99047c gl_rasterizer: Re-enable stream buffer memory due to global memory
Global memory is still using the stream buffer when it shouldn't. As a
temporary fix re-enable the stream buffer on compute.
2019-11-02 13:19:19 -03:00
ReinUsesLisp
38b6042bcb gl_rasterizer: Upload constant buffers with glNamedBufferSubData
Nvidia's OpenGL driver maps gl(Named)BufferSubData with some requirements
to a fast. This path has an extra memcpy but updates the buffer without
orphaning or waiting for previous calls. It can be seen as a better
model for "push constants" that can upload a whole UBO instead of 256
bytes.

This path has some requirements established here:
http://on-demand.gputechconf.com/gtc/2014/presentations/S4379-opengl-44-scene-rendering-techniques.pdf#page=24

Instead of using the stream buffer, this commits moves constant buffers
uploads to calls of glNamedBufferSubData and from my testing it brings a
performance improvement. This is disabled when the vendor is not Nvidia
since it brings performance regressions.
2019-11-02 05:05:34 -03:00
bunnei
daa6663ce0 Merge pull request #3046 from ReinUsesLisp/clean-gl-state
gl_state: Miscellaneous clean up
2019-10-29 22:50:04 -04:00
bunnei
4bc62dd758 Merge pull request #3035 from ReinUsesLisp/rasterizer-accelerated
rasterizer_accelerated: Add intermediary for GPU rasterizers
2019-10-29 22:06:41 -04:00
Rodrigo Locatti
efbd57f6e3 gl_state: Use std::array::fill instead of std::fill
Co-Authored-By: Mat M. <mathew1800@gmail.com>
2019-10-30 01:30:31 +00:00
ReinUsesLisp
b157ce5bbd gl_state: Move dirty checks to individual apply calls instead of Apply
This requires removing constness from some methods, but for consistency
it's removed in all methods.
2019-10-29 21:27:25 -03:00
ReinUsesLisp
385122dc97 gl_state: Remove ApplyDefaultState
OpenGL has defaults values we can trust. Remove these.
2019-10-29 21:27:25 -03:00
ReinUsesLisp
9651fece1f gl_state: Change SetDefaultViewports to use default constructor 2019-10-29 21:27:24 -03:00
ReinUsesLisp
3cb079e85f gl_state: Minor style changes 2019-10-29 21:27:24 -03:00
ReinUsesLisp
941e2036ca gl_state: Remove unused Citra TextureUnits 2019-10-29 21:27:24 -03:00
ReinUsesLisp
f15e087454 gl_state: Move initializers from constructor to class declaration 2019-10-29 21:27:23 -03:00
ReinUsesLisp
1589a146ed shader/node: Unpack bindless texture encoding
Bindless textures were using u64 to pack the buffer and offset from
where they come from. Drop this in favor of separated entries in the
struct.

Remove the usage of std::set in favor of std::list (it's not std::vector
to avoid reference invalidations) for samplers and images.
2019-10-29 20:53:48 -03:00
Rodrigo Locatti
e6c14ff73d Merge pull request #3004 from ReinUsesLisp/maxwell3d-cleanup
maxwell_3d: Remove unused entries
2019-10-29 23:46:33 +00:00