Commit Graph

362 Commits

Author SHA1 Message Date
bf2ddb8fd5 Merge pull request #3677 from FernandoS27/better-sync
Introduce Predictive Flushing and Improve ASYNC GPU
2020-04-22 22:09:38 -04:00
d9463f4562 vk_pipeline_cache: Fix unintentional memcpy into optional
The intention behind this was to assign a float to from an uint32_t, but
it was unintentionally being copied directly into the std::optional.

Copy to a temporary and assign that temporary to std::optional. This can
be replaced with std::bit_cast<float> once we are in C++20.
2020-04-22 21:36:05 -03:00
afae40a99e Merge pull request #3653 from ReinUsesLisp/nsight-aftermath
renderer_vulkan: Integrate Nvidia Nsight Aftermath on Windows
2020-04-22 11:39:01 -04:00
39e5b72948 Async GPU: Correct flushing behavior to be similar to old async GPU behavior. 2020-04-22 11:36:26 -04:00
644588fd88 ShaderCache/PipelineCache: Cache null shaders. 2020-04-22 11:36:25 -04:00
f616dc0b59 Address Feedback. 2020-04-22 11:36:24 -04:00
b752faf2d3 vk_fence_manager: Initial implementation 2020-04-22 11:36:19 -04:00
131b342130 OpenGL: Guarantee writes to Buffers. 2020-04-22 11:36:18 -04:00
1fb516cd97 GPU: Implement Flush Requests for Async mode. 2020-04-22 11:36:17 -04:00
b7bc3c2549 FenceManager: Manage syncpoints and rename fences to semaphores. 2020-04-22 11:36:16 -04:00
4adfc9bb08 Rasterizer: Document SignalFence & ReleaseFences and setup skeletons on Vulkan. 2020-04-22 11:36:14 -04:00
165ae823f5 ThreadManager: Sync async reads on accurate gpu. 2020-04-22 11:36:12 -04:00
8b1eb44b3e BufferCache: Implement OnCPUWrite and SyncGuestHost 2020-04-22 11:36:07 -04:00
da8f17715d GPU: Refactor synchronization on Async GPU 2020-04-22 11:36:06 -04:00
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
cb5b8ca886 Merge pull request #3733 from ambasta/patch-2
Initialize quad_indexed_pass before uint8_pass
2020-04-20 20:36:46 -04:00
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
5324b1d01e Initialize quad_indexed_pass before uint8_pass
Fixes Werror=reorder in gcc
2020-04-20 04:53:52 +05:30
85c17a2c35 Merge pull request #3694 from ReinUsesLisp/indexed-quads
vk_compute_pass: Implement indexed quads
2020-04-19 16:52:40 -04:00
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
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
b571c92dfd fixed_pipeline_state: Pack blending state
Reduce FixedPipelineState's size to 364 bytes.
2020-04-18 19:23:35 -03:00
548dd27f45 fixed_pipeline_state: Pack rasterizer state
Reduce FixedPipelineState's size to 600 bytes.
2020-04-18 19:22:57 -03:00
7790144a55 fixed_pipeline_state: Pack depth stencil state
Reduce FixedPipelineState's size to 632 bytes.
2020-04-18 19:22:11 -03:00
ab6704f20c fixed_pipeline_state: Pack attribute state
Reduce FixedPipelineState's size from 1384 to 664 bytes
2020-04-18 19:21:19 -03:00
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
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
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
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
11837e8f13 video_core: Amend doxygen comment references
Fixes broken documentation references.
2020-04-15 22:33:29 -04:00
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
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
3036067047 maxwell_to_vk: Add uint16 vertex formats 2020-04-15 04:06:30 -03:00
b4e43c64c8 maxwell_to_vk: Add missing breaks
Avoid invalid fallbacks.
2020-04-15 04:05:33 -03:00
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
37e5c4fa7c vk_rasterizer: Default to 1 viewports with a size of 0
Silence validation layer errors.
2020-04-14 04:44:34 -03:00
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
6cfe2a7246 renderer_vulkan: Remove Nvidia checkpoints 2020-04-13 17:33:59 -03:00
16105c6a66 renderer_vulkan: Catch device losses in more places 2020-04-13 17:33:59 -03:00
7e4a132a77 Merge pull request #3636 from ReinUsesLisp/drop-vk-hpp
renderer_vulkan: Drop Vulkan-Hpp
2020-04-13 17:08:04 -03:00
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
2905142f47 renderer_vulkan: Drop Vulkan-Hpp 2020-04-10 22:49:02 -03:00
51c6688e21 Merge pull request #3594 from ReinUsesLisp/vk-instance
yuzu: Drop SDL2 and Qt frontend Vulkan requirements
2020-04-10 20:06:55 -04:00
7cd6daf115 VkRasterizer: Eliminate Legacy code. 2020-04-08 18:59:09 -04:00
913f42a3a7 Memory: Address Feedback. 2020-04-08 13:40:46 -04:00
bf1d66b7c0 yuzu: Drop SDL2 and Qt frontend Vulkan requirements
Create Vulkan instances and surfaces from the Vulkan backend.
2020-04-07 16:32:19 -03:00
bc1b4b85b0 renderer_vulkan: Query device names from the backend 2020-04-07 02:23:23 -03:00
ea535d9470 Shader/Pipeline Cache: Use VAddr instead of physical memory for addressing. 2020-04-06 09:23:07 -04:00
3dd5c07454 Query Cache: Use VAddr instead of physical memory for adressing. 2020-04-06 09:23:07 -04:00
7fcd0fee6d Buffer Cache: Use vAddr instead of physical memory. 2020-04-06 09:23:06 -04:00