Commit Graph

106 Commits

Author SHA1 Message Date
ReinUsesLisp
33a0597603 vk_stream_buffer: Implement a stream buffer
This manages two kinds of streaming buffers: one for unified memory
models and one for dedicated GPUs. The first one skips the copy from the
staging buffer to the real buffer, since it creates an unified buffer.

This implementation waits for all fences to finish their operation
before "invalidating". This is suboptimal since it should allocate
another buffer or start searching from the beginning. There is room for
improvement here.

This could also handle AMD's "pinned" memory (a heap with 256 MiB) that
seems to be designed for buffer streaming.
2019-02-24 04:27:51 -03:00
ReinUsesLisp
f546fb35ed vk_scheduler: Implement a scheduler
The scheduler abstracts command buffer and fence management with an
interface that's able to do OpenGL-like operations on Vulkan command
buffers.

It returns by value a command buffer and fence that have to be used for
subsequent operations until Flush or Finish is executed, after that the
current execution context (the pair of command buffers and fences) gets
invalidated a new one must be fetched. Thankfully validation layers will
quickly detect if this is skipped throwing an error due to modifications
to a sent command buffer.
2019-02-22 01:33:32 -03:00
ReinUsesLisp
b675c97cdd vk_memory_manager: Implement memory manager
A memory manager object handles the memory allocations for a device. It
allocates chunks of Vulkan memory objects and then suballocates.
2019-02-19 03:42:28 -03:00
ReinUsesLisp
33a4cebc22 vk_resource_manager: Add VKResource interface
VKResource is an interface that gets signaled by a fence when it is free
to be reused.
2019-02-14 18:36:15 -03:00
bunnei
fcc3aa0bbf
Merge pull request #2113 from ReinUsesLisp/vulkan-base
vulkan: Add dependencies and device abstraction
2019-02-14 10:06:48 -05:00
ReinUsesLisp
8beca060d1 vk_device: Abstract device handling into a class
VKDevice contains all the data required to manage and initialize a
physical device. Its intention is to be passed across Vulkan objects to
query device-specific data (for example the logical device and the
dispatch loader).
2019-02-12 21:43:02 -03:00
ReinUsesLisp
18fe910957 renderer_vulkan: Add declarations file
This file is intended to be included instead of vulkan/vulkan.hpp. It
includes declarations of unique handlers using a dynamic dispatcher
instead of a static one (which would require linking to a Vulkan
library).
2019-02-12 18:33:02 -03:00
ReinUsesLisp
1ddcd0e6f0 kepler_compute: Fixup assert and rename engines
When I originally added the compute assert I used the wrong
documentation. This addresses that.

The dispatch register was tested with homebrew against hardware and is
triggered by some games (e.g. Super Mario Odyssey). What exactly is
missing to get a valid program bound by this engine requires more
investigation.
2019-02-10 19:29:33 -03:00
ReinUsesLisp
f087639e4a gl_shader_disk_cache: Compress GLSL code using LZ4 2019-02-06 22:23:39 -03:00
ReinUsesLisp
145c3ac89e gl_shader_disk_cache: Add file and move BaseBindings declaration 2019-02-06 22:20:57 -03:00
ReinUsesLisp
3b84e04af1 shader_decode: Implement LDG and basic cbuf tracking 2019-01-30 00:00:15 -03:00
ReinUsesLisp
af5d7e2c49 video_core: Rename glsl_decompiler to gl_shader_decompiler 2019-01-15 17:54:53 -03:00
ReinUsesLisp
a1b845b651 shader_decode: Implement VMAD and VSETP 2019-01-15 17:54:53 -03:00
ReinUsesLisp
a4f052f6b3 video_core: Replace gl_shader_decompiler 2019-01-15 17:54:50 -03:00
ReinUsesLisp
0c6fb456e0 glsl_decompiler: Implementation 2019-01-15 17:54:50 -03:00
ReinUsesLisp
15a0e1481d shader_ir: Initial implementation 2019-01-15 17:54:49 -03:00
ReinUsesLisp
3121408a90 gl_global_cache: Add dummy global cache manager 2019-01-08 17:47:45 -03:00
bunnei
abea6fa90c gpu: Rewrite GPU command list processing with DmaPusher class.
- More accurate impl., fixes Undertale (among other games).
2018-11-26 23:14:01 -05:00
ReinUsesLisp
dad3a6718e video_core: Move morton functions to their own file 2018-11-25 00:37:18 -03:00
Lioncash
9de523fd90 rasterizer_cache: Add missing virtual destructor to RasterizerCacheObject
Ensures that destruction will always do the right thing in any context.
2018-11-08 00:31:39 -05:00
Markus Wick
54df9fe29e gl_resource_manager: Split implementations in .cpp file.
Those implementations are quite costly, so there is no need to inline them to the caller.
Ressource deletion is often a performance bug, so in this way, we support to add breakpoints to them.
2018-11-06 14:40:39 +01:00
ReinUsesLisp
76754f5705 video_core: Move surface declarations out of gl_rasterizer_cache 2018-10-30 16:07:20 -03:00
ReinUsesLisp
80cbd81276 video_core: Move OpenGL specific utils to its renderer 2018-10-28 22:22:30 -03:00
ReinUsesLisp
3e2380327a gl_rasterizer: Implement quads topology 2018-10-04 00:03:44 -03:00
bunnei
fafc80d72e
Merge pull request #1290 from FernandoS27/shader-header
Implemented (Partialy) Shader Header
2018-09-17 18:53:14 -04:00
Subv
bb5eb4f20a GPU: Basic implementation of the Kepler Inline Memory engine (p2mf).
This engine writes data from a FIFO register into the configured address.
2018-09-12 13:57:08 -05:00
FernandoS27
e926757c8f Implemented (Partialy) Shader Header 2018-09-11 12:34:27 -04:00
Lioncash
8f4e09ba07 video_core/CMakeLists: Add missing gl_buffer_cache.h
Without this, the header file won't show up by default within IDEs such
as Visual Studio.
2018-09-06 14:49:51 -04:00
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
bunnei
b55d8111e6 renderer_opengl: Implement a new shader cache. 2018-08-27 18:26:46 -04:00
bunnei
382852418b video_core: Add RasterizerCache class for common cache management code. 2018-08-27 18:26:45 -04:00
bunnei
61779fa072 gl_rasterizer: Implement texture format ASTC_2D_4X4. 2018-06-18 01:56:59 -04:00
Subv
987a170665 GPU: Partially implemented the Maxwell DMA engine.
Only tiled->linear and linear->tiled copies that aren't offsetted are supported for now. Queries are not supported. Swizzled copies are not supported.
2018-06-12 11:27:36 -05:00
bunnei
459826a705 renderer_opengl: Add gl_shader_manager class. 2018-04-13 23:48:25 -04:00
bunnei
4e7e0f8112 shader_bytecode: Add initial module for shader decoding. 2018-04-13 23:48:19 -04:00
Subv
1ec8d2123d GPU: Implemented a gpu macro interpreter.
The Ryujinx macro interpreter and envydis were used as reference.

Macros are programs that are uploaded by the games during boot and can later be called by writing to their method id in a GPU command buffer.
2018-04-01 12:07:26 -05:00
bunnei
4369af6b7e maxwell_to_gl: Add module and function for decoding VertexType. 2018-03-26 21:16:55 -04:00
Subv
77fd0d47e7 Frontend: Ported the GPU breakpoints and surface viewer widgets from citra. 2018-03-24 11:31:49 -05:00
Subv
71ebc3e90d GPU: Preliminary work for texture decoding. 2018-03-24 11:30:56 -05:00
bunnei
6e3222363c renderer_gl: Port boilerplate rasterizer code over from Citra. 2018-03-20 00:07:32 -04:00
bunnei
d7b1ebe4a8 renderer_gl: Port over gl_shader_gen module from Citra. 2018-03-20 00:07:30 -04:00
bunnei
4bdb46e4c2 renderer_gl: Port over gl_shader_decompiler module from Citra. 2018-03-19 23:14:03 -04:00
bunnei
a3e10b1a72 renderer_gl: Port over gl_rasterizer_cache module from Citra. 2018-03-19 23:14:03 -04:00
bunnei
0e4b9cdde4 renderer_gl: Port over gl_stream_buffer module from Citra. 2018-03-19 23:14:02 -04:00
Subv
a64b936cbe GPU: Move the GPU's class constructor and destructors to a cpp file.
This should reduce recompile times when editing the Maxwell3D register structure.
2018-03-18 15:23:24 -05:00
Subv
6cddf9d88e Make a GPU class in VideoCore to contain the GPU state.
Also moved the GPU MemoryManager class to video_core since it makes more sense for it to be there.
2018-02-11 23:44:12 -05:00
Subv
e01a8f2187 GPU: Added a command processor to decode the GPU pushbuffers and forward the commands to their respective engines. 2018-02-11 22:42:48 -05:00
Lioncash
e710a1b989 CMakeLists: Derive the source directory grouping from targets themselves
Removes the need to store to separate SRC and HEADER variables, and then
construct the target in most cases.
2018-01-17 21:51:43 -05:00
James Rowe
1d28b2e142 Remove references to PICA and rasterizers in video_core 2018-01-12 19:11:03 -07:00
wwylele
0f35755572 pica/command_processor: build geometry pipeline and run geometry shader
The geometry pipeline manages data transfer between VS, GS and primitive assembler. It has known four modes:
 - no GS mode: sends VS output directly to the primitive assembler (what citra currently does)
 - GS mode 0: sends VS output to GS input registers, and sends GS output to primitive assembler
 - GS mode 1: sends VS output to GS uniform registers, and sends GS output to primitive assembler. It also takes an index from the index buffer at the beginning of each primitive for determine the primitive size.
 - GS mode 2: similar to mode 1, but doesn't take the index and uses a fixed primitive size.
hwtest shows that immediate mode also supports GS (at least for mode 0), so the geometry pipeline gets refactored into its own class for supporting both drawing mode.
In the immediate mode, some games don't set the pipeline registers to a valid value until the first attribute input, so a geometry pipeline reset flag is set in `pipeline.vs_default_attributes_setup.index` trigger, and the actual pipeline reconfigure is triggered in the first attribute input.
In the normal drawing mode with index buffer, the vertex cache is a little bit modified to support the geometry pipeline. Instead of OutputVertex, it now holds AttributeBuffer, which is the input to the geometry pipeline. The AttributeBuffer->OutputVertex conversion is done inside the pipeline vertex handler. The actual hardware vertex cache is believed to be implemented in a similar way (because this is the only way that makes sense).
Both geometry pipeline and GS unit rely on states preservation across drawing call, so they are put into the global state. In the future, the other three vertex shader units should be also placed in the global state, and a scheduler should be implemented on top of the four units. Note that the current gs_unit already allows running VS on it in the future.
2017-08-19 10:13:20 +03:00