Commit graph

2330 commits

Author SHA1 Message Date
Morph
78d80d99a0 maxwell_to_gl: Add 32 bit component sizes to (un)signed scaled formats
Add 32 bit component sizes to (un)signed scaled formats and group (un)signed normalized, scaled, and integer formats together.
2020-06-28 02:51:13 -04:00
Rodrigo Locatti
5872fc21fe
Merge pull request #4151 from ReinUsesLisp/gl-invalidations
gl_shader_cache: Avoid use after move for program size
2020-06-25 21:05:27 -03:00
David Marcec
a927d8be52 gl_device: Fix IsASTCSupported
Other targets were never actually checked
2020-06-25 19:12:56 +10:00
ReinUsesLisp
32a2dcd415 buffer_cache: Use buffer methods instead of cache virtual methods 2020-06-24 02:36:14 -03:00
ReinUsesLisp
39c97f1b65 gl_stream_buffer: Use InvalidateBufferData instead unmap and map
Making the stream buffer resident increases GPU usage significantly on
some games. This seems to be addressed invalidating the stream buffer
with InvalidateBufferData instead of using a Unmap + Map (with
invalidation flags).
2020-06-24 02:36:14 -03:00
ReinUsesLisp
41a4090320 gl_rasterizer: Use NV_vertex_buffer_unified_memory for vertex buffer robustness
Switch games are allowed to bind less data than what they use in a
vertex buffer, the expected behavior here is that these values are read
as zero. At the moment of writing this only D3D12, OpenGL and NVN through
NV_vertex_buffer_unified_memory support vertex buffer with a size limit.

In theory this could be emulated on Vulkan creating a new VkBuffer for
each (handle, offset, length) tuple and binding the expected data to it.
This is likely going to be slow and memory expensive when used on the
vertex buffer and we have to do it on all draws because we can't know
without analyzing indices when a game is going to read vertex data out
of bounds.

This is not a problem on OpenGL's BufferAddressRangeNV because it takes
a length parameter, unlike Vulkan's CmdBindVertexBuffers that only takes
buffers and offsets (the length is implicit in VkBuffer). It isn't a
problem on D3D12 either, because D3D12_VERTEX_BUFFER_VIEW on
IASetVertexBuffers takes SizeInBytes as a parameter (although I am not
familiar with robustness on D3D12).

Currently this only implements buffer ranges for vertex buffers,
although indices can also be affected. A KHR_robustness profile is not
created, but Nvidia's driver reads out of bound vertex data as zero
anyway, this might have to be changed in the future.

- Fixes SMO random triangles when capturing an enemy, getting hit, or
looking at the environment on certain maps.
2020-06-24 02:36:14 -03:00
ReinUsesLisp
32485917ba gl_buffer_cache: Mark buffers as resident
Make stream buffer and cached buffers as resident and query their
address. This allows us to use GPU addresses for several proprietary
Nvidia extensions.
2020-06-24 02:36:14 -03:00
ReinUsesLisp
73fb3a304b gl_device: Expose NV_vertex_buffer_unified_memory except on Turing
Expose NV_vertex_buffer_unified_memory when the driver supports it.

This commit adds a function the determine if a GL_RENDERER is a Turing
GPU. This is required because on Turing GPUs Nvidia's driver crashes
when the buffer is marked as resident or on DeleteBuffers. Without a
synchronous debug output (single threaded driver), it's likely that
the driver will crash in the first blocking call.
2020-06-24 02:36:14 -03:00
ReinUsesLisp
00c66a7289 gl_stream_buffer: Always use a non-coherent buffer 2020-06-24 02:35:33 -03:00
ReinUsesLisp
da79ec9565 gl_stream_buffer: Always use persistent memory maps
yuzu no longer supports platforms without persistent maps.
2020-06-24 02:35:33 -03:00
Rodrigo Locatti
b66ccaa376
Merge pull request #4129 from Morph1984/texture-shadow-lod-workaround
gl_shader_decompiler: Workaround textureLod when GL_EXT_texture_shadow_lod is not available
2020-06-24 01:51:15 -03:00
ReinUsesLisp
9f54cd4dad gl_shader_cache: Avoid use after move for program size
All programs had a size of zero due to this bug, skipping invalidations.

While we are at it, remove some unused forward declarations.
2020-06-23 22:54:42 -03:00
Morph
f77c897b8d gl_shader_decompiler: Enable GL_EXT_texture_shadow_lod if available
Enable GL_EXT_texture_shadow_lod if available. If this extension is not available, such as on Intel/AMD proprietary drivers, use textureGrad as a workaround.
2020-06-20 23:02:29 -04:00
Morph
1e65da971b gl_device: Check for GL_EXT_texture_shadow_lod 2020-06-20 22:14:32 -04:00
Lioncash
5865a10885 gl_arb_decompiler: Avoid several string copies
Variables that are marked as const cannot have the move constructor
invoked when returning from a function (the move constructor requires a
non-const variable so it can "steal" the resources from it.
2020-06-19 23:09:16 -04:00
Morph
8868fb745f maxwell_to_gl: Miscellaneous changes
maxwell_to_gl: Log unimplemented features under UNIMPLEMENTED_MSG instead of LOG_ERROR to bring into parity with maxwell_to_vk
maxwell_to_gl: Deduplicate logging in VertexType(), merging them into one.

maxwell_to_gl: Return GL_NEAREST instead of GL_LINEAR if an unknown texture filter mode is encountered.
maxwell_to_gl: Log the mipmap filter mode if an unknown value is passed in.
maxwell_to_gl: Reorder filtering modes to start with None, then Nearest, then Linear.
2020-06-18 04:56:31 -04:00
Rodrigo Locatti
edb2114bac
Merge pull request #4092 from Morph1984/image-bindings
gl_device: Reserve 4 image bindings for fragment stage
2020-06-18 04:59:48 -03:00
bunnei
798ec003ce
Merge pull request #4041 from ReinUsesLisp/arb-decomp
gl_arb_decompiler: Implement an assembly shader decompiler
2020-06-16 14:56:23 -04:00
Morph
e2f5d16540 gl_device: Reserve at least 4 image bindings for fragment stage
Due to the limitation of GL_MAX_IMAGE_UNITS being low (8) on Intel's and Nvidia's proprietary drivers, we have to reserve an appropriate amount of image bindings for each of the stages. So far games have been observed to use 4 image bindings on the fragment stage (Kirby Star Allies) and 1 on the vertex stage (TWD series).
No games thus far in my limited testing used more than 4 images concurrently and across all currently active programs.
This fixes shader compilation errors on Kirby Star Allies on OpenGL (GLSL/GLASM)
2020-06-16 03:03:07 -04:00
Rodrigo Locatti
0bd9bc7201
Merge pull request #4066 from ReinUsesLisp/shared-ptr-buf
buffer_cache: Avoid passing references of shared pointers and misc style changes
2020-06-15 22:29:32 -03:00
bunnei
92021a344c
Merge pull request #4064 from ReinUsesLisp/invalidate-buffers
gl_rasterizer: Mark vertex buffers as dirty after buffer cache invalidation
2020-06-14 00:29:16 -04:00
bunnei
c2ea1e1bcb
Merge pull request #4049 from ReinUsesLisp/separate-samplers
shader/texture: Join separate image and sampler pairs offline
2020-06-13 13:48:27 -04:00
bunnei
5633887569
Merge pull request #3986 from ReinUsesLisp/shader-cache
shader_cache: Implement a generic runtime shader cache
2020-06-12 23:14:48 -04:00
ReinUsesLisp
87011a97f9 gl_arb_decompiler: Implement FSwizzleAdd 2020-06-11 22:12:07 -03:00
ReinUsesLisp
a63a0daa5e gl_arb_decompiler: Implement an assembly shader decompiler
Emit code compatible with NV_gpu_program5.
This should emit code compatible with Fermi, but it wasn't tested on
that architecture. Pascal has some issues not present on Turing GPUs.
2020-06-11 22:12:07 -03:00
bunnei
83e3b77ed7
Merge pull request #4027 from ReinUsesLisp/3d-slices
texture_cache: Implement rendering to 3D textures
2020-06-09 21:52:15 -04:00
ReinUsesLisp
6508cdd003 buffer_cache: Avoid passing references of shared pointers and misc style changes
Instead of using as template argument a shared pointer, use the
underlying type and manage shared pointers explicitly. This can make
removing shared pointers from the cache more easy.

While we are at it, make some misc style changes and general
improvements (like insert_or_assign instead of operator[] + operator=).
2020-06-09 18:30:49 -03:00
ReinUsesLisp
7646f2c21d gl_rasterizer: Mark vertex buffers as dirty after buffer cache invalidation
Vertex buffers bindings become invalid after the stream buffer is
invalidated. We were originally doing this, but it got lost at some
point.

- Fixes Animal Crossing: New Horizons, but it affects everything.
2020-06-08 20:24:16 -03:00
bunnei
3626254f48
Merge pull request #4040 from ReinUsesLisp/nv-transform-feedback
gl_rasterizer: Use NV_transform_feedback for XFB on assembly shaders
2020-06-08 16:18:33 -04:00
bunnei
98d2461529
Merge pull request #4052 from ReinUsesLisp/debug-output
renderer_opengl: Only enable DEBUG_OUTPUT when graphics debugging is enabled
2020-06-08 10:16:41 -04:00
ReinUsesLisp
3c2ae53b4c texture_cache: Handle 3D texture blits with one layer 2020-06-08 05:01:00 -03:00
ReinUsesLisp
c95c254f3e texture_cache: Implement rendering to 3D textures
This allows rendering to 3D textures with more than one slice.
Applications are allowed to render to more than one slice of a texture
using gl_Layer from a VTG shader.

This also requires reworking how 3D texture collisions are handled, for
now, this commit allows rendering to slices but not to miplevels. When a
render target attempts to write to a mipmap, we fallback to the previous
implementation (copying or flushing as needed).

- Fixes color correction 3D textures on UE4 games (rainbow effects).
- Allows Xenoblade games to render to 3D textures directly.
2020-06-08 05:01:00 -03:00
ReinUsesLisp
abcea1bb18 rasterizer_cache: Remove files and includes
The rasterizer cache is no longer used. Each cache has its own generic
implementation optimized for the cached data.
2020-06-07 04:32:57 -03:00
ReinUsesLisp
678f95e4f8 vk_pipeline_cache: Use generic shader cache
Trivial port the generic shader cache to Vulkan.
2020-06-07 04:32:57 -03:00
ReinUsesLisp
b96f65b62b gl_shader_cache: Use generic shader cache
Trivially port the generic shader cache to OpenGL.
2020-06-07 04:32:57 -03:00
ReinUsesLisp
e78d681a6c gl_device: Black list NVIDIA 443.24 for fast buffer uploads
Skip fast buffer uploads on Nvidia 443.24 Vulkan beta driver on OpenGL.
This driver throws the following error when calling BufferSubData or
BufferData on buffers that are candidates for fast constant buffer
uploads. This is the equivalens to push constants on Vulkan, except that
they can access the full buffer. The error:

Unknown internal debug message. The NVIDIA OpenGL driver has encountered
an out of memory error. This application might
behave inconsistently and fail.

If this error persists on future drivers, we might have to look deeper
into this issue. For now, we can black list it and log it as a temporary
solution.
2020-06-06 02:56:42 -03:00
ReinUsesLisp
354fbe701e renderer_opengl: Only enable DEBUG_OUTPUT when graphics debugging is enabled
Avoids logging when it's not relevant. This can potentially reduce
driver's internal thread overhead.
2020-06-05 21:21:12 -03:00
ReinUsesLisp
5b2b6d594c shader/texture: Join separate image and sampler pairs offline
Games using D3D idioms can join images and samplers when a shader
executes, instead of baking them into a combined sampler image. This is
also possible on Vulkan.

One approach to this solution would be to use separate samplers on
Vulkan and leave this unimplemented on OpenGL, but we can't do this
because there's no consistent way of determining which constant buffer
holds a sampler and which one an image. We could in theory find the
first bit and if it's in the TIC area, it's an image; but this falls
apart when an image or sampler handle use an index of zero.

The used approach is to track for a LOP.OR operation (this is done at an
IR level, not at an ISA level), track again the constant buffers used as
source and store this pair. Then, outside of shader execution, join
the sample and image pair with a bitwise or operation.

This approach won't work on games that truly use separate samplers in a
meaningful way. For example, pooling textures in a 2D array and
determining at runtime what sampler to use.

This invalidates OpenGL's disk shader cache :)

- Used mostly by D3D ports to Switch
2020-06-05 00:24:51 -03:00
bunnei
22369df357
Merge pull request #4031 from Morph1984/fix-gs-outputs
gl_shader_decompiler: Fix geometry shader outputs on Intel drivers
2020-06-04 15:18:51 -04:00
ReinUsesLisp
3d99b449d3 gl_rasterizer: Use NV_transform_feedback for XFB on assembly shaders
NV_transform_feedback, NV_transform_feedback2 and
ARB_transform_feedback3 with NV_transform_feedback interactions allows
implementing transform feedbacks as dynamic state.

Maxwell implements transform feedbacks as dynamic state, so using these
extensions with TransformFeedbackStreamAttribsNV allows us to properly
emulate transform feedbacks without having to recompile shaders when the
state changes.
2020-06-03 20:22:12 -03:00
bunnei
623b93a2b3
Merge pull request #4014 from ReinUsesLisp/astc-nvidia
gl_device: Avoid devices with CAVEAT_SUPPORT on ASTC
2020-06-02 17:43:33 -04:00
bunnei
597d8b4bd4
Merge pull request #4006 from ReinUsesLisp/squash-ubos
glsl: Squash constant buffers into a single SSBO when we hit the limit
2020-06-02 14:58:50 -04:00
Morph
74f2e5f1a4 gl_shader_decompiler: Declare gl_Layer and gl_ViewportIndex within gl_PerVertex for vertex and tessellation shaders 2020-06-01 15:35:44 -04:00
Morph
70188d69b0 gl_shader_decompiler: Fix geometry shader outputs for Intel drivers
On Intel's proprietary drivers, gl_Layer and gl_ViewportIndex are not allowed members of gl_PerVertex block, causing the shader to fail to compile. Fix this by declaring these variables outside of gl_PerVertex.
2020-06-01 15:34:05 -04:00
bunnei
6c0b1a9ee2
Merge pull request #3996 from ReinUsesLisp/front-faces
fixed_pipeline_state,gl_rasterizer: Swap negative viewport checks for front faces
2020-06-01 14:04:35 -04:00
ReinUsesLisp
0ee310ebdc gl_device: Avoid devices with CAVEAT_SUPPORT on ASTC
This avoids using Nvidia's ASTC decoder on OpenGL.
The last time it was profiled, it was slower than yuzu's decoder.

While we are at it, fix a bug in the texture cache when native ASTC is
not supported.
2020-05-31 21:34:34 -03:00
ReinUsesLisp
ee21e4ecd3 glsl: Squash constant buffers into a single SSBO when we hit the limit
Avoids compilation errors at the cost of shader build times and runtime
performance when a game hits the limit of uniform buffers we can use.
2020-05-31 21:33:49 -03:00
bunnei
edbf3144d2
Merge pull request #3958 from FernandoS27/gl-debug
OpenGL: Enable Debug Context and Synchronous debugging when graphics debugging is enabled
2020-05-31 17:04:27 -04:00
Morph
bb8ef38152 gl_device: Enable compute shaders for Intel proprietary drivers
Previously we were disabling compute shaders on Intel's proprietary driver due to broken compute. This has been fixed in the latest Intel drivers. Re-enable compute for Intel proprietary drivers and remove the check for broken compute.
2020-05-31 03:21:07 -04:00
bunnei
058ec22787
Merge pull request #3982 from ReinUsesLisp/membar-cts
shader/other: Implement MEMBAR.CTS
2020-05-30 11:51:42 -04:00