Commit graph

4291 commits

Author SHA1 Message Date
Fernando Sahmkow
a60a22d9c2 Texture Cache: Implement OnCPUWrite and SyncGuestHost 2020-04-22 11:36:05 -04:00
Fernando Sahmkow
084ceb925a UI: Replasce accurate GPU option for GPU Accuracy Level 2020-04-22 11:36:04 -04:00
ReinUsesLisp
6f47bd9641 vk_memory_manager: Remove unified memory model flag
All drivers (even Intel) seem to have a device local memory type that is
not host visible. Remove this flag so all devices follow the same path.

This fixes a crash when trying to map to host device local memory on
integrated devices.
2020-04-21 22:06:38 -03:00
bunnei
d64290884a
Merge pull request #3714 from lioncash/copies
gl_shader_decompiler: Avoid copies where applicable
2020-04-21 20:16:02 -04:00
ReinUsesLisp
488ed8bd02 vk_rasterizer: Add lazy default buffer maker and use it for empty buffers
Introduce a default buffer getter that lazily constructs an empty
buffer. This is intended to match OpenGL's buffer 0.

Use this for disabled vertex and uniform buffers.

While we are at it, include vertex buffer usages for staging buffers to
silence validation errors.
2020-04-21 19:55:52 -03:00
ReinUsesLisp
0bbae63300 gl_rasterizer: Fix buffers without size
On NVN buffers can be enabled but have no size. According to deko3d and
the behavior we see in Animal Crossing: New Horizons these buffers get
the special address of 0x1000 and limit themselves to 0xfff.

Implement buffers without a size by binding a null buffer to OpenGL
without a side.

1d1930beea/source/maxwell/gpu_3d_vbo.cpp (L62-L63)
2020-04-21 19:55:44 -03:00
Rodrigo Locatti
f293b15611
Merge pull request #3718 from ReinUsesLisp/better-pipeline-state
fixed_pipeline_state: Pack structure, use memcmp and CityHash on it
2020-04-21 18:17:58 -03:00
bunnei
9bf3abcb63
Merge pull request #3698 from lioncash/warning
General: Resolve minor assorted warnings
2020-04-21 14:11:18 -04:00
bunnei
d3e0cefa60
Merge pull request #3695 from ReinUsesLisp/default-attributes
maxwell_3d: Initialize format attributes constant as one
2020-04-20 21:40:18 -04:00
ReinUsesLisp
8734ccb0cb shader/arithmetic_integer: Fix LEA_IMM encoding
The operand order in LEA_IMM was flipped compared to nvdisasm. Fix that
using nxas as reference:

8dbc389957/table.h (L122)
2020-04-20 21:54:59 -03:00
Mat M
cb5b8ca886
Merge pull request #3733 from ambasta/patch-2
Initialize quad_indexed_pass before uint8_pass
2020-04-20 20:36:46 -04:00
Fernando Sahmkow
ec2f8f4272
Merge pull request #3700 from ReinUsesLisp/stream-buffer-sizes
vk_stream_buffer: Fix out of memory on boot on recent Nvidia drivers
2020-04-20 09:37:42 -04:00
Amit Prakash Ambasta
5324b1d01e
Initialize quad_indexed_pass before uint8_pass
Fixes Werror=reorder in gcc
2020-04-20 04:53:52 +05:30
Rodrigo Locatti
4932010c6f
Merge pull request #3729 from lioncash/globals
dma_pusher: Remove reliance on the global system instance
2020-04-19 19:12:40 -03:00
bunnei
85c17a2c35
Merge pull request #3694 from ReinUsesLisp/indexed-quads
vk_compute_pass: Implement indexed quads
2020-04-19 16:52:40 -04:00
Lioncash
44e959157b dma_pusher: Remove reliance on the global system instance
With this, the video core is now has no calls to the global system
instance at all.
2020-04-19 16:12:08 -04:00
bunnei
2ea7a70da0
Merge pull request #3686 from lioncash/table
texture_cache/format_lookup_table: Fix incorrect green, blue, and alpha indices
2020-04-19 15:33:33 -04:00
bunnei
73db83c0ab
Merge pull request #3679 from lioncash/track
track: Eliminate redundant copies
2020-04-19 01:22:47 -04:00
Jan Beich
afcc84a172 renderer_vulkan: assume X11 if not Windows/macOS after bf1d66b7c0
Render.Vulkan <Error> video_core/renderer_vulkan/renderer_vulkan.cpp:CreateInstance:131: Presentation not supported on this platform
Render.Vulkan <Error> video_core/renderer_vulkan/renderer_vulkan.cpp:CreateSurface:378: Presentation not supported on this platform
Core <Critical> core/core.cpp:Load:199: Failed to initialize system (Error 5)!
2020-04-19 00:32:23 +00:00
ReinUsesLisp
c81bf06d03 vulkan/wrapper: Sort physical devices
Sort discrete GPUs over the rest, Nvidia over AMD, AMD over Intel, Intel
over the rest. This gives us a somewhat consistent order when Optimus
is removed (renderdoc does this when it's attached).

This can break the configuration of users with an Intel GPU that
manually remove Optimus on yuzu. That said, it's a very unlikely to
happen.
2020-04-18 21:31:15 -03:00
ReinUsesLisp
d62f57cf5a fixed_pipeline_state: Hash and compare the whole structure
Pad FixedPipelineState's size to 384 bytes to be a multiple of 16.

Compare the whole struct with std::memcmp and hash with CityHash. Using
CityHash instead of a naive hash should reduce the number of collisions.
Improve used type traits to ensure this operation is safe.

With these changes the improvements to the hashable pipeline state are:

Optimized structure
Hash:            89 ns
Comparison:     103 ns
Construction*:  164 ns
Struct size:    384 bytes

Original structure
Hash:           148 ns
Equal:          174 ns
Construction*:  281 ns
Size:          1384 bytes

* Attribute state initialization is not measured

These measures are averages taken with std::chrono::high_accuracy_clock
on MSVC shipped on Visual Studio 16.6.0 Preview 2.1.
2020-04-18 19:57:26 -03:00
ReinUsesLisp
b571c92dfd fixed_pipeline_state: Pack blending state
Reduce FixedPipelineState's size to 364 bytes.
2020-04-18 19:23:35 -03:00
ReinUsesLisp
548dd27f45 fixed_pipeline_state: Pack rasterizer state
Reduce FixedPipelineState's size to 600 bytes.
2020-04-18 19:22:57 -03:00
ReinUsesLisp
7790144a55 fixed_pipeline_state: Pack depth stencil state
Reduce FixedPipelineState's size to 632 bytes.
2020-04-18 19:22:11 -03:00
ReinUsesLisp
ab6704f20c fixed_pipeline_state: Pack attribute state
Reduce FixedPipelineState's size from 1384 to 664 bytes
2020-04-18 19:21:19 -03:00
Mat M
5305806071
Merge pull request #3716 from bunnei/fix-another-impl-fallthrough
video_core: gl_shader_decompiler: Fix implicit fallthrough errors.
2020-04-18 15:17:52 -04:00
bunnei
03726fb7f5 video_core: gl_shader_decompiler: Fix implicit fallthrough errors. 2020-04-18 15:15:21 -04:00
Lioncash
bf328ed35a gl_shader_decompiler: Avoid copies where applicable
Avoids unnecessary reference count increments where applicable and also
avoids reallocating a vector.

Unlikely to make a huge difference, but given how trivial of an
amendment it is, why not?
2020-04-17 20:48:52 -04:00
Markus Wick
07fbef1776 video_code: Fix implicit switch fallthrough.
Since yesterday, this breaks the build on linux.
So let's fix it.
2020-04-17 23:43:35 +02:00
ReinUsesLisp
a7b6bd56d7 vk_stream_buffer: Fix out of memory on boot on recent Nvidia drivers
Nvidia recently introduced a new memory type for data streaming
(awesome!), but yuzu was assuming that all heaps had enough memory
for the assumed stream buffer size (256 MiB).

This worked fine on AMD but Nvidia's new memory heap was smaller than
256 MiB. This commit changes this assumption and allocates a bit less
than the size of the preferred heap, with a maximum of 256 MiB (to avoid
allocating all system memory on integrated devices).

- Fixes a crash on NVIDIA 450.82.0.0
2020-04-17 18:12:48 -03:00
Rodrigo Locatti
990c0b184f
Revert "gl_shader_cache: Use CompileDepth::FullDecompile on GLSL" 2020-04-17 17:41:48 -03:00
bunnei
b8f5c71f2d
Merge pull request #3666 from bunnei/new-vmm
Implement a new virtual memory manager
2020-04-17 16:33:08 -04:00
bunnei
ca3af2961c
Merge pull request #3682 from lioncash/uam
gl_query_cache: Resolve use-after-move in CachedQuery move assignment operator
2020-04-17 01:24:08 -04:00
bunnei
32fc2aae3c video_core: memory_manager: Updates for Common::PageTable changes. 2020-04-17 00:59:34 -04:00
bunnei
4caff51710 core: memory: Move to Core::Memory namespace.
- helpful to disambiguate Kernel::Memory namespace.
2020-04-17 00:59:28 -04:00
Lioncash
e2d8be1ca2 General: Resolve warnings related to missing declarations 2020-04-16 23:43:34 -04:00
Lioncash
678ac54749 decode/memory: Resolve unused variable warning
Only the first element of the returned pair is ever used.
2020-04-16 22:45:44 -04:00
Lioncash
d159643fd7 decode/texture: Resolve unused variable warnings.
Some variables aren't used, so we can remove these.

Unfortunately, diagnostics are still reported on structured bindings
even when annotated with [[maybe_unused]], so we need to unpack the
elements that we want to use manually.
2020-04-16 22:45:41 -04:00
Lioncash
f522abd8ab decode/texture: Collapse loop down into std::generate
Same behavior, less code.
2020-04-16 22:29:07 -04:00
Lioncash
7e2d60de26 decode/texture: Eliminate trivial missing field initializer warnings
We can just specify the initializers.
2020-04-16 22:27:21 -04:00
bunnei
79c1269f0f
Merge pull request #3673 from lioncash/extra
CMakeLists: Specify -Wextra on linux builds
2020-04-16 21:12:33 -04:00
ReinUsesLisp
238c6016f9 maxwell_3d: Initialize format attributes constant as one
nouveau expects this to be true but it doesn't set it.
2020-04-16 21:15:07 -03:00
ReinUsesLisp
c961770900 vk_compute_pass: Implement indexed quads
Implement indexed quads (GL_QUADS used with glDrawElements*) with a
compute pass conversion.

The compute shader converts from uint8/uint16/uint32 indices to uint32.
The format is passed through push constants to avoid having different
variants of the same shader.

- Used by Fast RMX
- Used by Xenoblade Chronicles 2 (it still has graphical due to
synchronization issues on Vulkan)
2020-04-16 21:12:32 -03:00
Fernando Sahmkow
c81f256111
Merge pull request #3600 from ReinUsesLisp/no-pointer-buf-cache
buffer_cache: Return handles instead of pointer to handles
2020-04-16 19:58:13 -04:00
ReinUsesLisp
090fd3fefa buffer_cache: Return handles instead of pointer to handles
The original idea of returning pointers is that handles can be moved.
The problem is that the implementation didn't take that in mind and made
everything harder to work with. This commit drops pointer to handles and
returns the handles themselves. While it is still true that handles can
be invalidated, this way we get an old handle instead of a dangling
pointer.

This problem can be solved in the future with sparse buffers.
2020-04-16 02:33:34 -03:00
Rodrigo Locatti
a5a2ee8766
Merge pull request #3689 from lioncash/unused-var
decode/shift: Remove unused variable within Shift()
2020-04-16 02:05:54 -03:00
Rodrigo Locatti
d196ce0f71
Merge pull request #3688 from lioncash/nequal
surface_view: Add missing operator!= to ViewParams
2020-04-16 01:39:51 -03:00
Rodrigo Locatti
4209dba1f6
Merge pull request #3680 from lioncash/static
gl_device: Mark stage_swizzle as constexpr
2020-04-16 01:26:23 -03:00
Rodrigo Locatti
60e8de7c95
Merge pull request #3687 from lioncash/constness
surface_base: Make IsInside() a const member function
2020-04-16 01:22:50 -03:00
Rodrigo Locatti
612966399b
Merge pull request #3685 from lioncash/copies
control_flow: Make use of std::move in TryInspectAddress()
2020-04-16 01:22:40 -03:00
Lioncash
cd2a12e78f decode/shift: Remove unused variable within Shift()
Removes a redundant variable that is already satisfied by the IsFull()
utility function.
2020-04-16 00:16:06 -04:00
Lioncash
5fbe8785d2 surface_view: Add missing operator!= to ViewParams
Provides logical symmetry to the interface.
2020-04-16 00:03:12 -04:00
Lioncash
d551c910bb surface_base: Make IsInside() a const member function
This doesn't modify internal state, so this can be made const.
2020-04-15 23:59:35 -04:00
bunnei
319df1db77
Merge pull request #3683 from lioncash/docs
video_core: Amend doxygen comment references
2020-04-15 23:54:58 -04:00
Lioncash
636c8ab85b texture_cache/format_lookup_table: Fix incorrect green, blue, and alpha indices
Previously these were all using the red component to derive the indices,
which is definitely not intentional.
2020-04-15 23:50:46 -04:00
Lioncash
72a224d3fc control_flow: Make use of std::move in TryInspectAddress()
Eliminates redundant atomic reference count increments and decrements.
2020-04-15 23:31:22 -04:00
Lioncash
11837e8f13 video_core: Amend doxygen comment references
Fixes broken documentation references.
2020-04-15 22:33:29 -04:00
Lioncash
3a60f19eaf gl_query_cache: Resolve use-after-move in CachedQuery move assignment operator
Avoids potential invalid junk data from being read.
2020-04-15 22:20:06 -04:00
Lioncash
71fb156611 gl_device: Mark stage_swizzle as constexpr
Previously this was mutable even though it shouldn't be.
2020-04-15 21:59:13 -04:00
Lioncash
e15ec2705c track: Eliminate redundant copies
Two variables can be references, while two others can be std::moved.
Makes for 4 less atomic reference count increments and decrements.
2020-04-15 21:50:09 -04:00
Lioncash
1c340c6efa CMakeLists: Specify -Wextra on linux builds
Allows reporting more cases where logic errors may exist, such as
implicit fallthrough cases, etc.

We currently ignore unused parameters, since we currently have many
cases where this is intentional (virtual interfaces).

While we're at it, we can also tidy up any existing code that causes
warnings. This also uncovered a few bugs as well.
2020-04-15 21:33:46 -04:00
Rodrigo Locatti
65cbb122ea
Merge pull request #3649 from FernandoS27/3d-fix
Texture Cache: Read current data when flushing a 3D segment.
2020-04-15 17:06:55 -03:00
Fernando Sahmkow
e33196d4e7
Merge pull request #3612 from ReinUsesLisp/red
shader/memory: Implement RED.E.ADD and minor changes to ATOM
2020-04-15 15:03:49 -04:00
Lioncash
213fff67bc CMakeLists: Make -Wreorder a compile-time error
This can result in silent logic bugs within code, and given the amount
of times these kind of warnings are caused, they should be flagged at
compile-time so no new code is submitted with them.
2020-04-15 14:14:41 -04:00
Mat M
64b5985f0a
Merge pull request #3662 from ReinUsesLisp/constant-attrs
gl_rasterizer: Implement constant vertex attributes
2020-04-15 11:54:50 -04:00
Fernando Sahmkow
6789d88a9c Texture Cache: Read current data when flushing a 3D segment.
This PR corrects flushing of 3D segments when data of other segments is
mixed, this aims to preserve the data in place.
2020-04-15 11:46:17 -04:00
Mat M
9208d555b7
Merge pull request #3668 from ReinUsesLisp/vtx-format-16ui
maxwell_to_vk: Add uint16 vertex formats
2020-04-15 11:43:52 -04:00
Mat M
ab72696beb
Merge pull request #3656 from ReinUsesLisp/glsl-full-decompile
gl_shader_cache: Use CompileDepth::FullDecompile on GLSL
2020-04-15 03:17:46 -04:00
Mat M
4878d6bb49
Merge pull request #3654 from ReinUsesLisp/fix-fb-attach
gl_texture_cache: Fix layered texture attachment base level
2020-04-15 03:17:18 -04:00
Mat M
50c0a92db8
Merge pull request #3663 from ReinUsesLisp/fcmp-rc
shader/arithmetic: Add FCMP_CR variant
2020-04-15 03:16:56 -04:00
Mat M
13331a3a32
Merge pull request #3664 from ReinUsesLisp/fe3h-black-squares
Revert "gl_shader_decompiler: Implement merges with bitfieldInsert"
2020-04-15 03:14:28 -04:00
ReinUsesLisp
3036067047 maxwell_to_vk: Add uint16 vertex formats 2020-04-15 04:06:30 -03:00
ReinUsesLisp
b4e43c64c8 maxwell_to_vk: Add missing breaks
Avoid invalid fallbacks.
2020-04-15 04:05:33 -03:00
ReinUsesLisp
0ca456830f vk_blit_screen: Initialize all members in VkPipelineViewportStateCreateInfo
When the dynamic state is specified, pViewports and pScissors are
ignored, quoting the specification:

  pViewports is a pointer to an array of VkViewport structures, defining
  the viewport transforms. If the viewport state is dynamic, this member
  is ignored.

That said, AMD's proprietary driver itself seem to read it regardless of
what the specification says.
2020-04-15 03:30:08 -03:00
Rodrigo Locatti
0b132e8cc1
Merge pull request #3657 from ReinUsesLisp/viewport-zero
vk_rasterizer: Default to 1 viewports with a size of 0
2020-04-15 01:51:17 -03:00
Fernando Sahmkow
daddbeffd1
Texture Cache: Only do buffer copies on accurate GPU. (#3634)
This is a simple optimization as Buffer Copies are mostly used for texture recycling. They are, however, useful when games abuse undefined behavior but most 3D APIs forbid it.
2020-04-14 23:21:00 -04:00
ReinUsesLisp
fd6371eba7 Revert "gl_shader_decompiler: Implement merges with bitfieldInsert"
This reverts commit 05cf270836.

Apparently the first approach using floats instead of bitfieldInert
worked better for Fire Emblem: Three Houses. Reverting to get that
behavior back.
2020-04-14 21:24:33 -03:00
ReinUsesLisp
fefe7f18f9 shader/arithmetic: Add FCMP_CR variant
Adds another variant of FCMP.
2020-04-14 19:11:04 -03:00
ReinUsesLisp
6dfcabc800 gl_rasterizer: Implement constant vertex attributes
Credits go to gdkchan from Ryujinx for finding constant attributes are
used in retail games.
2020-04-14 17:58:53 -03:00
ReinUsesLisp
37e5c4fa7c vk_rasterizer: Default to 1 viewports with a size of 0
Silence validation layer errors.
2020-04-14 04:44:34 -03:00
ReinUsesLisp
453d7419d9 gl_shader_cache: Use CompileDepth::FullDecompile on GLSL
From my testing on a Splatoon 2 shader that takes 3800ms on average to
compile changing to FullDecompile reduces it to 900ms on average.

The shader decoder will automatically fallback to a more naive method if
it can't use full decompile.
2020-04-14 01:34:20 -03:00
ReinUsesLisp
0e232cfdc1 renderer_vulkan: Integrate Nvidia Nsight Aftermath on Windows
Adds optional support for Nsight Aftermath. It is enabled through
ENABLE_NSIGHT_AFTERMATH in cmake. A path to the SDK has to be provided
by the environment variable NSIGHT_AFTERMATH_SDK.

Nsight Aftermath allows an application to generate "minidumps" of the
GPU state when a device loss happens. By analysing these on Nsight we
can know what a game was doing and why it triggered a device loss.

The dump is generated inside %APPDATA%\yuzu\log\gpucrash and this
directory is deleted every time a new instance is initialized with
Nsight enabled.

To enable it on yuzu there has a to be a driver and device capable of
running Nsight Aftermath on Vulkan. That means only Turing based GPUs
on the latest stable driver, beta drivers won't work for now.

It is manually enabled in Configuration>Debug>Enable Graphics Debugging
because when using all debugging capabilities there is a runtime cost.
2020-04-14 00:39:21 -03:00
ReinUsesLisp
21dc842171 gl_texture_cache: Fix layered texture attachment base level
The base level is already included in the texture view. If we specify
the base level in the texture again, this will end up in the incorrect
level and potentially out of bounds.
2020-04-13 18:24:56 -03:00
ReinUsesLisp
6cfe2a7246 renderer_vulkan: Remove Nvidia checkpoints 2020-04-13 17:33:59 -03:00
ReinUsesLisp
16105c6a66 renderer_vulkan: Catch device losses in more places 2020-04-13 17:33:59 -03:00
Rodrigo Locatti
7e4a132a77
Merge pull request #3636 from ReinUsesLisp/drop-vk-hpp
renderer_vulkan: Drop Vulkan-Hpp
2020-04-13 17:08:04 -03:00
Mat M
fbf13d3f48
Merge pull request #3651 from ReinUsesLisp/line-widths
gl_rasterizer: Implement line widths and smooth lines
2020-04-13 10:19:59 -04:00
Mat M
08266d70ba
Merge pull request #3638 from ReinUsesLisp/remove-preserve-contents
texture_cache: Remove preserve_contents
2020-04-13 10:19:01 -04:00
Mat M
c4001225f6
Merge pull request #3631 from ReinUsesLisp/more-astc
texture/astc: More small ASTC optimizations
2020-04-13 10:17:32 -04:00
Mat M
7b62212461
Merge pull request #3619 from ReinUsesLisp/i2i
shader/conversion: Implement I2I sign extension, saturation and selection
2020-04-13 10:17:07 -04:00
Mat M
3351e1e94f
Merge pull request #3627 from ReinUsesLisp/layered-view
gl_texture_cache: Attach view instead of base texture for layered attchments
2020-04-13 10:16:18 -04:00
Mat M
d37d899431
Merge pull request #3646 from ReinUsesLisp/fix-glsl-turing
gl_shader_decompiler: Improve generated code in HMergeH*
2020-04-13 10:15:12 -04:00
Mat M
47036859eb
Merge pull request #3633 from ReinUsesLisp/clean-texdec
shader/texture: Remove type mismatches management from shader decoder
2020-04-13 10:13:05 -04:00
ReinUsesLisp
76615b9f34 gl_rasterizer: Implement line widths and smooth lines
Implements "legacy" features from OpenGL present on hardware such as
smooth lines and line width.
2020-04-13 01:30:34 -03:00
ReinUsesLisp
05cf270836 gl_shader_decompiler: Implement merges with bitfieldInsert
This also fixes Turing issues but it avoids doing more bitcasts. This
should improve the generated code while also avoiding more points where
compilers can flush floats.
2020-04-12 22:39:59 -03:00
Fernando Sahmkow
3d91dbb21d
Merge pull request #3578 from ReinUsesLisp/vmnmx
shader/video: Partially implement VMNMX
2020-04-12 10:44:03 -04:00
ReinUsesLisp
75eb953575 gl_shader_decompiler: Improve generated code in HMergeH*
Avoiding bitwise expressions, this fixes Turing issues in shaders using
half float merges that affected several games.
2020-04-12 05:06:55 -03:00
ReinUsesLisp
76f178ba6e shader/video: Partially implement VMNMX
Implements the common usages for VMNMX. Inputs with a different size
than 32 bits are not supported and sign mismatches aren't supported
either.

VMNMX works as follows:
It grabs Ra and Rb and applies a maximum/minimum on them (this is
defined by .MX), having in mind the input sign. This result can then be
saturated. After the intermediate result is calculated, it applies
another operation on it using Rc. These operations are merges,
accumulations or another min/max pass.

This instruction allows to implement with a more flexible approach GCN's
min3 and max3 instructions (for instance).
2020-04-12 00:34:42 -03:00
ReinUsesLisp
a7baf6fee4 video_core: Add MSAA registers in 3D engine and TIC
This adds the registers used for multisampling. It doesn't implement
anything for now.
2020-04-12 00:21:27 -03:00
ReinUsesLisp
94b0e2e5da texture_cache: Remove preserve_contents
preserve_contents was always true. We can't assume we don't have to
preserve clears because scissored and color masked clears exist.

This removes preserve_contents and assumes it as true at all times.
2020-04-11 01:51:02 -03:00