Markus Wick
d3ad9469a1
gl_rasterizer: Implement a VAO cache.
...
This patch caches VAO objects instead of re-emiting all pointers per draw call.
Configuring this pointers is known as a fast task, but it yields too many GL
calls. So for better performance, just bind the VAO instead of 16 pointers.
2018-09-05 18:46:35 +02:00
Markus Wick
50a806ea67
renderer_opengl: Implement a buffer cache.
...
The idea of this cache is to avoid redundant uploads. So we are going
to cache the uploaded buffers within the stream_buffer and just reuse
the old pointers.
The next step is to implement a VBO cache on GPU memory, but for now,
I want to check the overhead of the cache management. Fetching the
buffer over PCI-E should be quite fast.
2018-09-05 08:03:50 +02:00
Markus Wick
99a71580c4
gl_shader_cache: Use an u32 for the binding point cache.
...
The std::string generation with its malloc and free requirement
was a noticeable overhead. Also switch to an ordered_map to
avoid the std::hash call. As those maps usually have a size of
two elements, the lookup time shall not matter.
2018-09-04 21:04:41 +02:00
Markus Wick
10bc725944
Update microprofile scopes.
...
Blame the subsystems which deserve the blame :)
The updated list is not complete, just the ones I've spotted on random sampling the stack trace.
2018-09-04 11:04:26 +02:00
bunnei
7f7eb29323
gl_rasterizer_cache: Use accurate framebuffer setting for accurate copies.
2018-08-31 13:07:28 -04:00
bunnei
d647d9550c
gl_renderer: Cache textures, framebuffers, and shaders based on CPU address.
2018-08-31 13:07:27 -04:00
bunnei
16d65182f9
gl_rasterizer: Fix issues with the rasterizer cache.
...
- Use a single cached page map.
- Fix calculation of ending page.
2018-08-31 13:07:27 -04:00
bunnei
4d7e1662c8
Merge pull request #1193 from lioncash/priv
...
gpu: Make memory_manager private
2018-08-28 12:28:57 -04:00
Lioncash
45fb74d262
gpu: Make memory_manager private
...
Makes the class interface consistent and provides accessors for
obtaining a reference to the memory manager instance.
Given we also return references, this makes our more flimsy uses of
const apparent, given const doesn't propagate through pointers in the
way one would typically expect. This makes our mutable state more
apparent in some places.
2018-08-28 11:11:50 -04:00
Lioncash
6771a18c6c
gl_rasterizer: Remove unused variables
2018-08-28 10:46:29 -04:00
bunnei
b55d8111e6
renderer_opengl: Implement a new shader cache.
2018-08-27 18:26:46 -04:00
bunnei
a0e1566dc5
gl_rasterizer_cache: Update to use RasterizerCache base class.
2018-08-27 18:26:46 -04:00
Lioncash
8fd9eb71b4
gl_rasterizer: Correct assertion condition in SyncLogicOpState()
...
Previously the assert would always be hit, since it was the equivalent
of: array == nullptr, which is never true.
2018-08-23 23:00:54 -04:00
bunnei
a4ac3bed6c
gl_rasterizer: Implement stencil test.
...
- Used by Splatoon 2.
2018-08-23 11:08:49 -04:00
bunnei
da3da6be90
gl_rasterizer: Implement partial color clear and stencil clear.
2018-08-23 11:08:48 -04:00
bunnei
c7f2fb2151
Merge pull request #1157 from lioncash/vec
...
gl_shader_gen: Use a std::vector to represent program code instead of std::array
2018-08-23 02:19:00 -04:00
Lioncash
1fd979f50a
gl_shader_gen: Use a std::vector to represent program code instead of std::array
...
While convenient as a std::array, it's also quite a large set of data as
well (32KB). It being an array also means data cannot be std::moved. Any
situation where the code is being set or relocated means that a full
copy of that 32KB data must be done.
If we use a std::vector we do need to allocate on the heap, however, it
does allow us to std::move the data we have within the std::vector into
another std::vector instance, eliminating the need to always copy the
program data (as std::move in this case would just transfer the pointers
and bare necessities over to the new vector instance).
2018-08-22 17:04:44 -04:00
Lioncash
dd35b4b18a
renderer_opengl: Namespace OpenGL code
...
Namespaces all OpenGL code under the OpenGL namespace.
Prevents polluting the global namespace and allows clear distinction
between other renderers' code in the future.
2018-08-22 06:14:47 -04:00
bunnei
125d7122ac
Merge pull request #1124 from Subv/logic_ops
...
GPU: Implemented logic ops.
2018-08-22 01:05:25 -04:00
bunnei
bf89a99839
Merge pull request #1123 from lioncash/screen
...
rasterizer_interface: Remove renderer-specific ScreenInfo type from AccelerateDraw() in RasterizerInterface
2018-08-21 01:18:34 -04:00
Subv
2b9eee4d1e
GPU: Implemented the logic op functionality of the GPU.
...
This will ASSERT if blending is enabled at the same time as logic ops.
2018-08-20 18:44:47 -05:00
Lioncash
46ef072cf9
rasterizer_interface: Remove ScreenInfo from AccelerateDraw()'s signature
...
This is an OpenGL renderer-specific data type. Given that, this type
shouldn't be used within the base interface for the rasterizer. Instead,
we can pass this information to the rasterizer via reference.
2018-08-20 19:43:05 -04:00
Subv
3fe77be392
Rasterizer: Don't attempt to copy over the old texture's data when doing a format reinterpretation if we're only going to clear the framebuffer.
2018-08-20 15:20:35 -05:00
Subv
e0f66c1fbf
GLRasterizer: Implemented instanced vertex arrays.
...
Before each draw call, for every enabled vertex array configured as instanced, we take the current instance id and divide it by its configured divisor, then we multiply that by the corresponding stride and increment the start address by the resulting amount. This way we can simulate the vertex array being incremented once per instance without actually using OpenGL's instancing functions.
2018-08-18 14:42:26 -05:00
Subv
c5284efd4f
Rasterizer: Implemented instanced rendering.
...
We keep track of the current instance and update an uniform in the shaders to let them know which instance they are.
Instanced vertex arrays are not yet implemented.
2018-08-14 22:25:07 -05:00
bunnei
290439a6a5
gl_rasterizer: Fix upload size for constant buffers.
2018-08-14 20:44:19 -04:00
Markus Wick
0eb39922f6
gl_rasterizer: Use a shared helper to upload from CPU memory.
2018-08-12 16:10:26 +02:00
Markus Wick
0af7e93763
gl_state: Don't track constant buffer mappings.
2018-08-12 16:10:26 +02:00
Markus Wick
6ff7906ddc
gl_rasterizer: Use the stream buffer for constant buffers.
2018-08-12 16:10:26 +02:00
Markus Wick
ce722e317b
gl_rasterizer: Use the streaming buffer itself for the constant buffer.
...
Don't emut copies, especially not for data, which is used once. They just end in a huge GPU overhead.
2018-08-12 15:48:59 +02:00
Markus Wick
6f6bba3ff1
gl_rasterizer: Use a helper for aligning the buffer.
2018-08-12 15:47:35 +02:00
Markus Wick
d7298ec262
Update the stream_buffer helper from Citra.
...
Please see https://github.com/citra-emu/citra/pull/3666 for more details.
2018-08-12 15:47:35 +02:00
bunnei
ee07041b3a
Merge pull request #1020 from lioncash/namespace
...
core: Namespace EmuWindow
2018-08-11 22:40:08 -04:00
bunnei
9c977d2215
Merge pull request #1021 from lioncash/warn
...
gl_rasterizer: Silence implicit truncation warning in SetupShaders()
2018-08-11 22:39:46 -04:00
Subv
2dad1204e8
RasterizerGL: Ignore invalid/unset vertex attributes.
...
This should make the es2gears example not crash anymore.
2018-08-11 20:36:40 -05:00
Lioncash
28e90fa0e0
gl_rasterizer: Silence implicit truncation warning in SetupShaders()
...
Previously this would warn of truncating a std::size_t to a u32. This is
safe because we'll obviously never have more than UINT32_MAX amount of
uniform buffers.
2018-08-11 20:32:03 -04:00
Lioncash
0a93b45b6a
core: Namespace EmuWindow
...
Gets the class out of the global namespace.
2018-08-11 20:20:21 -04:00
bunnei
57626fda7b
Merge pull request #1004 from lioncash/unused
...
gl_rasterizer_cache: Remove unused viewport parameter of GetFramebufferSurfaces()
2018-08-10 12:13:32 -04:00
bunnei
2156cb3cbe
Revert "gl_state: Temporarily disable culling and depth test."
2018-08-10 10:39:46 -04:00
Lioncash
0e1510ac29
gl_rasterizer_cache: Remove unused viewport parameter of GetFramebufferSurfaces()
2018-08-09 20:55:41 -04:00
bunnei
f765a6b902
Merge pull request #984 from bunnei/rt-none
...
gl_rasterizer: Do not render when no render target is configured.
2018-08-09 13:12:28 -04:00
bunnei
ddec200290
gl_rasterizer: Do not render when no render target is configured.
...
- Used by Super Mario Odyssey.
2018-08-08 19:29:45 -04:00
bunnei
9ceceb212f
Merge pull request #981 from bunnei/cbuf-corrupt
...
maxwell_3d: Use correct const buffer size and check bounds.
2018-08-08 19:16:34 -04:00
bunnei
57982df105
maxwell_3d: Use correct const buffer size and check bounds.
...
- Fixes mem corruption with Super Mario Odyssey and Pokkén Tournament DX.
2018-08-08 02:10:25 -04:00
bunnei
8c6338b6f9
renderer_opengl: Use trace log in a few places.
2018-08-08 01:53:23 -04:00
David Marcec
b96010bfa9
added braces for conditions
2018-08-05 11:36:55 +10:00
David Marcec
6d1e30e041
fix the attrib format for ints
2018-08-05 11:29:21 +10:00
Lioncash
0f2ac928f2
video_core: Make global EmuWindow instance part of the base renderer class
...
Makes the global a member of the RendererBase class. We also change this
to be a reference. Passing any form of null pointer to these functions
is incorrect entirely, especially given the code itself assumes that the
pointer would always be in a valid state.
This also makes it easier to follow the lifecycle of instances being
used, as we explicitly interact the renderer with the rasterizer, rather
than it just operating on a global pointer.
2018-08-01 21:40:30 -04:00
Lioncash
0162f8b3a7
gl_rasterizer: Replace magic number with GL_INVALID_INDEX in SetupConstBuffers()
...
This is just the named constant that OpenGL provides, so we can use that
instead of using a literal -1
2018-07-24 12:24:49 -04:00
Lioncash
16139ed53b
gl_rasterizer: Use std::string_view instead of std::string when checking for extensions
...
We can avoid heap allocations here by just using a std::string_view
instead of performing unnecessary copying of the string data.
2018-07-24 12:10:37 -04:00
Lioncash
b5eb3905cd
gl_rasterizer: Use in-class member initializers where applicable
...
We can just assign to the members directly in these cases.
2018-07-24 12:08:12 -04:00
bunnei
69c45ce71c
gl_rasterizer: Implement texture border color.
2018-07-23 23:34:42 -04:00
bunnei
4301f0b539
gl_rasterizer_cache: Use GPUVAddr as cache key, not parameter set.
2018-07-21 21:51:06 -04:00
bunnei
d8c60029d6
gl_rasterizer: Use zeta_enable register to enable depth buffer.
2018-07-21 21:51:06 -04:00
Lioncash
863579736c
gpu: Rename Get3DEngine() to Maxwell3D()
...
This makes it match its const qualified equivalent.
2018-07-20 18:34:49 -04:00
bunnei
cf30c4be22
gl_state: Temporarily disable culling and depth test.
2018-07-18 23:21:43 -04:00
Lioncash
3a4841e403
core: Don't construct instance of Core::System, just to access its live instance
...
This would result in a lot of allocations and related object
construction, just to toss it all away immediately after the call.
These are definitely not intentional, and it was intended that all of
these should have been accessing the static function GetInstance()
through the name itself, not constructed instances.
2018-07-18 18:18:27 -04:00
bunnei
2cb3fdca86
Merge pull request #598 from bunnei/makedonecurrent
...
OpenGL: Use MakeCurrent/DoneCurrent for multithreaded rendering.
2018-07-14 20:18:11 -07:00
bunnei
05cb10530f
OpenGL: Use MakeCurrent/DoneCurrent for multithreaded rendering.
2018-07-14 02:50:35 -04:00
Subv
b37354cca8
GPU: Always enable the depth write when clearing the depth buffer.
...
The GPU ignores that register when clearing, but OpenGL obeys the glDepthMask parameter, so we set the depth mask to GL_TRUE when clearing the depth buffer. It will be restored to the correct value automatically on the next draw call.
2018-07-14 00:52:23 -05:00
bunnei
8aeff9cf8e
gl_rasterizer: Fix check for if a shader stage is enabled.
2018-07-12 22:57:57 -04:00
bunnei
c4015cd93a
gl_shader_gen: Implement dual vertex shader mode.
...
- When VertexA shader stage is enabled, we combine with VertexB program to make a single Vertex Shader stage.
2018-07-12 22:25:36 -04:00
bunnei
854f474f52
gl_rasterizer: Flip triangles when regs.viewport_transform[0].scale_y is negative.
...
- Fixes a regression with Binding of Isaac.
2018-07-08 16:16:24 -04:00
bunnei
8b815877a6
Merge pull request #622 from Subv/unused_tex
...
GPU: Ignore unused textures and corrected the TEX shader instruction decoding.
2018-07-05 11:29:17 -04:00
Subv
4bda9693be
GPU: Ignore textures that the GLSL compiler deemed unused when binding textures to the shaders.
2018-07-04 15:20:12 -05:00
Subv
c1bebdef5e
GPU: Flip the triangle front face winding if the GPU is configured to not flip the triangles.
...
OpenGL's default behavior is already correct when the GPU is configured to flip the triangles.
This fixes 1-2 Switch's splash screen.
2018-07-04 10:26:46 -05:00
Subv
5a9df3c675
GPU: Only configure the used framebuffers during clear.
...
Don't try to configure the color buffer if it is not being cleared, it may not be completely valid at this point.
2018-07-03 22:32:59 -05:00
bunnei
c996787d84
Merge pull request #609 from Subv/clear_buffers
...
GPU: Implemented the CLEAR_BUFFERS register.
2018-07-03 19:34:34 -04:00
Subv
78443a7f29
GPU: Factor out the framebuffer configuration code for both Clear and Draw commands.
2018-07-03 16:56:47 -05:00
Subv
c1811ed3d1
GPU: Support clears that don't clear the color buffer.
2018-07-03 16:56:47 -05:00
Subv
be51120d23
GPU: Bind and clear the render target when the CLEAR_BUFFERS register is written to.
2018-07-03 16:56:44 -05:00
bunnei
15e68cdbaa
Merge pull request #607 from jroweboy/logging
...
Logging - Customizable backends
2018-07-03 00:26:45 -04:00
bunnei
e3ca561ea0
Merge pull request #612 from bunnei/fix-cull
...
gl_rasterizer: Only set cull mode and front face if enabled.
2018-07-02 23:48:52 -04:00
bunnei
a9cacd03f6
gl_rasterizer: Only set cull mode and front face if enabled.
2018-07-02 22:22:25 -04:00
Subv
65c664560c
GPU: Don't try to parse the depth test function if the depth test is disabled.
2018-07-02 21:02:46 -05:00
James Rowe
0d46f0df12
Update clang format
2018-07-02 21:45:47 -04:00
James Rowe
638956aa81
Rename logging macro back to LOG_*
2018-07-02 21:45:47 -04:00
bunnei
92c7135065
Merge pull request #608 from Subv/depth
...
GPU: Implemented the depth buffer and depth test + culling
2018-07-02 21:24:43 -04:00
Subv
a6d4903aaf
GPU: Set up the culling configuration on each draw.
2018-07-02 19:51:29 -05:00
Sebastian Valle
055f1546d7
Merge pull request #606 from Subv/base_vertex
...
GPU: Fixed the index offset and implement BaseVertex when doing indexed rendering.
2018-07-02 14:07:38 -05:00
Subv
18c8ae7750
GPU: Set up the depth test state on every draw.
2018-07-02 13:33:06 -05:00
Subv
0f929762b3
GPU: Implemented the Z24S8 depth format and load the depth framebuffer.
2018-07-02 12:42:04 -05:00
Subv
4c59105adf
GPU: Implement offsetted rendering when using non-indexed drawing.
2018-07-02 11:23:36 -05:00
Subv
fca3d1cc65
GPU: Fixed the index offset rendering, and implemented the base vertex functionality.
...
This fixes Stardew Valley.
2018-07-02 11:22:17 -05:00
Subv
80c5e8ae99
GPU: Ignore disabled textures and textures with an invalid address.
2018-07-02 09:43:38 -05:00
bunnei
1dd754590f
gl_rasterizer_cache: Implement caching for texture and framebuffer surfaces.
...
gl_rasterizer_cache: Improved cache management based on Citra's implementation.
gl_surface_cache: Add some docstrings.
2018-06-27 00:15:44 -04:00
bunnei
3f9f047375
gl_rasterizer: Implement AccelerateDisplay to forward textures to framebuffers.
2018-06-27 00:08:03 -04:00
bunnei
5f57ab1b2a
gl_rasterizer_cache: Remove Citra's rasterizer cache, always load/flush surfaces.
2018-06-27 00:08:03 -04:00
bunnei
10422f3c18
gl_rasterizer: Workaround for when exceeding max UBO size.
2018-06-26 23:07:34 -04:00
bunnei
8447d20a11
gl_state: Fix state management for texture swizzle.
2018-06-26 17:15:58 -04:00
bunnei
36dedae842
Merge pull request #554 from Subv/constbuffer_ubo
...
Rasterizer: Use UBOs instead of SSBOs for uploading const buffers.
2018-06-26 10:25:56 -04:00
Subv
a3d82ef5d9
Build: Fixed some MSVC warnings in various parts of the code.
2018-06-20 11:39:10 -05:00
Jules Blok
7c7f4a9be2
gl_rasterizer: Get loose on independent blending.
2018-06-18 09:27:06 +02:00
Subv
2a7653142d
Rasterizer: Use UBOs instead of SSBOs for uploading const buffers.
...
This should help a bit with GPU performance once we're GPU-bound.
2018-06-09 18:02:05 -05:00
Subv
c011b6f67e
GPU: Synchronize the blend state on every draw call.
...
Only independent blending on render target 0 is implemented for now.
This fixes the elongated squids in Splatoon 2's boot screen.
2018-06-08 17:05:52 -05:00
Subv
c712dafaee
GPU: Added registers for normal and independent blending.
2018-06-08 17:04:41 -05:00
Subv
47dc5e0dab
Rasterizer: Flush the written region when writing shader uniform data before copying it to the uniform buffers.
...
This fixes the flip_viewport uniform having invalid values when drawing.
2018-06-08 12:22:39 -05:00
bunnei
ee318d4015
Merge pull request #543 from Subv/uniforms
...
GLRenderer: Write the shader stage configuration UBO data *before* copying it to the GPU.
2018-06-07 11:21:36 -04:00
Subv
86146ef819
GLRenderer: Write the shader stage configuration UBO data *before* copying it to the GPU.
...
This should fix the bug with the vs_config UBO being uninitialized during shader execution.
2018-06-07 08:33:23 -05:00
Subv
47629c89a8
GPU: Support changing the texture swizzles for Maxwell textures.
2018-06-06 18:36:15 -05:00