Commit graph

1196 commits

Author SHA1 Message Date
tgsm
d6c530d08c video_core: use nested namespaces 2019-02-19 03:09:57 -05:00
Weiyi Wang
f409342ab5
Merge pull request #4613 from BreadFish64/gles5
video_core: add GLES support
2019-02-17 15:44:39 -05:00
Weiyi Wang
de1128c60d
Merge pull request #4628 from FearlessTobi/backport-some-stuff
Backport various minor frontend review changes from yuzu and port minor PR from yuzu
2019-02-14 11:45:27 -05:00
Weiyi Wang
4b2397aa77
Merge pull request #4633 from BreadFish64/fully-invalid
video_core: improve efficiency of CachedSurface::IsSurfaceFullyInvalid
2019-02-14 11:44:46 -05:00
ReinUsesLisp
fc6e33d812 video_core: Sort predicate order to have semantically positive value first 2019-02-13 21:06:37 +01:00
SachinVin
d63acfc1e9 video_core: add workarounds to enable GLES support
video_core: shorten GetGLSLVersionString

video_core: make GLES version and extensions consistent

video_core: move some logic to LoadShader

video_core: deduplicate fragment shader precision specifier
2019-02-12 16:57:18 -06:00
BreadFish64
d90f733330 video_core: improve efficiency of CachedSurface::IsSurfaceFullyInvalid 2019-02-09 17:33:18 -06:00
fearlessTobi
6be1b4d293 renderer_base: backport minor changes from yuzu 2019-02-06 17:16:27 +01:00
Weiyi Wang
b5f2318ae7 gl_rasterizer: change shadow_texture_bias from shader config var to shader uniform
Games can frequently change this register. Using it as shader config var would generates a lot of shaders
2019-02-02 20:40:08 -05:00
Lioncash
d2aac218a6 renderer_opengl: Correct forward declaration of FramebufferLayout
This is actually a struct, not a class, which can lead to compilation
warnings.
2018-12-27 02:48:05 +01:00
Weiyi Wang
07d6d90bb3 gl_rasterizer_cache/MortonCopy: avoid read/write to invalid address 2018-12-09 13:20:44 -05:00
Weiyi Wang
7e8ba6ed8e Memory: move memory chunk into pImpl and make them dynamically allocated
Otherwise MSVC would give out-of-memory error on compile time
2018-12-06 13:30:47 -05:00
Weiyi Wang
d18cda5a5d Memory: move MarkRegionCached into class 2018-12-05 20:21:14 -05:00
Weiyi Wang
296c458e0e Memory: move GetPhysicalPointer and IsValidPhysicalAddress into class 2018-12-05 20:21:14 -05:00
Weiyi Wang
8bb404c82a VideoCore: pass in memory system
currently set as global. Will change when videocore itself is wrapped as a class
2018-12-05 20:16:42 -05:00
Francois Berder
c598fd7228 renderer_opengl: Fix string comparison
Signed-off-by: Francois Berder <fberder@outlook.fr>
2018-12-04 22:12:35 +00:00
zhupengfei
071b41cb61
citra_qt, video_core: Screenshot functionality
Allows capturing screenshot at the current internal resolution (native for software renderer), but a setting is available to capture it in other resolutions. The screenshot is saved to a single PNG in the current layout.
2018-11-30 14:14:41 +08:00
Markus Wick
62036bdea8 gl_rasterizer_cache: Add profiles for Copy and Blit
They were missed, and Copy is very high in profile here. It doesn't block the GPU,
but it stalls the driver thread. So with our bad GL instructions, this might block quite a while.
2018-11-17 15:11:56 +01:00
Markus Wick
2b793797e0 gl_resource_manager: Profile creation and deletion 2018-11-17 15:07:30 +01:00
Markus Wick
79696a1a43 gl_stream_buffer: Profile orphaning of stream buffer.
This serialize to the driver thread and so it may block for a while.
So if it is in the benchmark, we get noticed if it happens too often.
2018-11-17 15:03:20 +01:00
Markus Wick
392547a97c 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.
2018-11-17 15:01:40 +01:00
Tobias
46e8237e7e Port yuzu-emu/yuzu#1137: "renderer_opengl: Namespace OpenGL code" (#4423)
* 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.

* Also namespace TextureCubeConfig
2018-11-16 23:29:10 -08:00
Weiyi Wang
9458e4d8ec CoreTiming: wrap into class 2018-11-04 10:26:38 -05:00
B3n30
2306af3600 Handle cases when std::optional does not contain a value 2018-10-05 16:51:33 +02:00
B3n30
d37a2270d6 Replace boost::optional with std::optional where possible 2018-10-05 13:51:09 +02:00
Lioncash
333b6a556a gl_stream_buffer: Fix use of bitwise OR instead of logical OR in Map()
This was very likely intended to be a logical OR based off the
conditioning and testing of inversion in one case.

Even if this was intentional, this is the kind of non-obvious thing one
should be clarifying with a comment.
2018-09-21 17:20:42 -05:00
Weiyi Wang
7d8f115185 Prefix all size_t with std::
done automatically by executing regex replace `([^:0-9a-zA-Z_])size_t([^0-9a-zA-Z_])` -> `$1std::size_t$2`
2018-09-06 16:03:28 -04:00
Weiyi Wang
12bba80d48
Merge pull request #4174 from wwylele/proctex-round-fix
gl_rasiterzer/proctex: revert back to round() for Nearest sampling
2018-09-04 01:36:16 +03:00
fearlessTobi
488694d01f Replace ToBool() with static_cast() 2018-09-01 15:50:57 +02:00
Weiyi Wang
99f4ea999c gl_rasiterzer/proctex: revert back to round() for Nearest sampling
This change to floor() was made in 2927c88, which was a result of doing some hwtest. It turned out that it was buggy edge cases in PICA, and for most cases round() still applies
2018-09-01 09:11:15 -04:00
Weiyi Wang
ffd78b635e
Merge pull request #4163 from FearlessTobi/port-1097
Port #1097 from yuzu: "renderer_opengl: Treat OpenGL errors as critical."
2018-08-31 21:43:52 +03:00
bunnei
ffdc27351a renderer_opengl: Treat OpenGL errors as critical. 2018-08-31 01:56:38 +02:00
Lioncash
6d280cb49a gl_stream_buffer: Add missing header guard
Prevents potential compilation errors from occuring due to multiple
inclusions
2018-08-31 01:45:44 +02:00
fearlessTobi
f61c9c3eb7 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.
2018-08-25 15:20:40 +02:00
Merry
0a5621fafc
Merge pull request #3970 from FearlessTobi/more-popup-madness
citra_qt: Add more verbose popups for video_core errors
2018-08-24 19:21:35 +01:00
Lioncash
d944d1120f video_core: Remove unimplemented Start() function prototype
Given this has no definition, we can just remove it entirely.
2018-08-23 16:58:30 +02:00
fearlessTobi
d1c5f01afe Add more verbose popups for video_core errors 2018-08-19 15:48:40 +02:00
zhupengfei
0c37140690 video_core: Allow changing background color while emulation is running
As @jroweboy stated, this is just one more atomic in VideoCore.
2018-08-04 15:11:51 +08:00
James Rowe
14b0435df2
Merge pull request #3994 from FearlessTobi/replace-clamp-functions
Remove MathUtil::Clamp and replace it with its std:: counterpart
2018-08-02 11:08:07 -06:00
James Rowe
4b59c1b601
Merge pull request #4042 from wwylele/i-love-drivers
gl_rasterizer: apply AMD hack only when the vendor is AMD
2018-08-02 11:01:25 -06:00
wwylele
266f2b6242 gl_rasterizer: only apply AMD hack when the vendor is AMD 2018-07-31 22:58:17 +03:00
Valentin Vanelslande
b2ad88dac1
swrasterizer/lighting: remove newline in logging macro
Software rasterizer version of #3904
2018-07-30 16:09:16 -05:00
fearlessTobi
15abf35889 Address feedback by wwylele 2018-07-30 14:46:13 +02:00
fearlessTobi
71e1d6e25d Fix compile errors 2018-07-26 13:23:25 +02:00
fearlessTobi
7a3e126a4f Replace MathUtil::Clamp with its std counterpart 2018-07-24 19:08:17 +02:00
Weiyi Wang
78685065cf
Merge pull request #3916 from wwylele/mipmap-proctex
gl_rasterizer: implement mipmap for procedural texture
2018-07-19 11:51:57 +03:00
wwylele
431fe444a9 gl_rasterizer: call glTextureBarrier when an image is bound to both texture and framebuffer 2018-07-09 23:11:29 +03:00
Weiyi Wang
7c7adc64fd
Merge pull request #3898 from wwylele/shut-up-texture5
swrasterizer, gl_shader_gen: return 0.0 for Disabled texture unit 0
2018-07-07 18:45:13 +03:00
wwylele
2927c88fd3 gl_rasterizer: implement mipmap for proctex 2018-07-06 01:13:02 +03:00
Valentin Vanelslande
b12942a231
gl_shader_gen: remove newline in logging macro 2018-06-30 19:15:15 -05:00
Valentin Vanelslande
fef231dc5b
gl_shader_manager: fix macro 2018-06-30 11:24:38 -05:00
wwylele
4d4b833a00 swrasterizer, gl_shader_gen: return 0.0 for Disabled texture unit 0 2018-06-30 11:36:39 +03:00
wwylele
0eab948728 reformat all files with clang-format 2018-06-29 16:56:12 +03:00
wwylele
7c5a76e58b log: replace all NGLOG with LOG 2018-06-29 14:18:07 +03:00
NarcolepticK
9ae70e733f video-core: Migrate logging macros (#3878)
* video-core: Migrate logging macros

* video-core: Fixed missed clang format

* video-core: Migrated LOG_GENERIC macro
2018-06-29 00:13:30 +03:00
Weiyi Wang
80f6510355
Merge pull request #3851 from wwylele/shut-up-texture3
gl_shader_gen: lower log level of using disabled proctex
2018-06-28 18:03:05 +03:00
Markus Wick
c4ff0ba137 renderer_opengl: Renumber all texture units.
We spend lots of texture units for our texture buffers. As they are now feed from
one buffer, there is no need to have the big gap in the list of IDs.
2018-06-22 17:21:28 +02:00
Markus Wick
46f18d6800 gl_shader_gen: Inline now constant texture buffer. 2018-06-22 17:20:55 +02:00
Markus Wick
0838c87dac gl_stream_buffer: Only flush the host buffer if anything was written.
This might happen in the new TBO upload path.
2018-06-22 17:20:55 +02:00
Markus Wick
831d4f9aeb gl_rasterizer: Use the shared texture buffer for the proctex lut. 2018-06-22 17:20:55 +02:00
Markus Wick
1ca6d2ea8d gl_rasterizer: Use the shared texture buffer for the noise, color and alpha map. 2018-06-22 17:20:55 +02:00
Markus Wick
63fb7dcc1b gl_rasterizer: Use the shared texture buffer for the fog lut. 2018-06-22 17:20:55 +02:00
Markus Wick
4679487640 gl_rasterizer: Use the shared texture buffer for the lighting lut. 2018-06-22 17:20:54 +02:00
Markus Wick
5a9cde138d gl_rasterizer: Add a new dirty flag for any lighting lut. 2018-06-22 17:17:48 +02:00
Markus Wick
a8396cdbed renderer_opengl: Create shared texture buffer. 2018-06-22 17:05:40 +02:00
Markus Wick
298ebe3752 gl_rasterizer: Inline texture buffer uploads. 2018-06-22 17:04:47 +02:00
Markus Wick
8a8c6f059f gl_rasterizer: Move TBO syncing helper to one function. 2018-06-22 17:04:47 +02:00
Markus Wick
10fba63b94 gl_rasterizer: Provide texture buffer offsets.
This allows us to move all data into one TBO.
2018-06-22 17:03:38 +02:00
Weiyi Wang
f50e505a5a
Merge pull request #3778 from wwylele/shadow-hw-image-load-store
gl_rasterizer: implement shadow map 2D/Cube - Image load/store version
2018-06-22 17:47:09 +03:00
wwylele
2f9b7bdfa9 gl_shader_gen: lower log level of using disabled proctex 2018-06-20 18:34:01 +03:00
James Rowe
574824a26c
Merge pull request #3632 from N00byKing/3dtv_botenable
Add Support for Stereoscopic 3D
2018-06-17 21:05:15 -06:00
James Rowe
cf9bfe0690
Merge pull request #3787 from wwylele/shader-jit-state
shader/jit: preserve integer & condition register across invocation
2018-06-09 18:38:05 -06:00
James Rowe
2dac1a9590
Merge pull request #3788 from wwylele/shader-jit-breakc
shader/jit: implement breakc
2018-06-09 18:36:46 -06:00
N00byKing
523c52c708 renderer_opengl: Add Universal 3D Layout Adaption 2018-06-01 18:24:26 +02:00
jmorriz124
8c0ede544f 3dtv botenable improved (#1)
* Fixed crash when right eye isn't available

* Enabled swap screens in stereo views.  Fixed window alignment in stereo
views to handle all screen aspect ratios.

* Minor code cleanup and clang fomat updates.

* Minor cleanup of swapped and aspect ratio code
2018-06-01 17:05:29 +02:00
N00byKing
2814bbc3da renderer_opengl: Allow usage of Stereoscopic 3D 2018-06-01 17:01:06 +02:00
wwylele
781912e854 gl_rasterize: implement shadow mapping using image load/store 2018-06-01 14:26:44 +03:00
Weiyi Wang
08b119153d
Merge pull request #3799 from wwylele/sigh
gl_rasterizer: reset texture state context after every draw
2018-06-01 14:24:28 +03:00
wwylele
9060e08e49 shader/jit: implement breakc 2018-06-01 13:04:39 +03:00
wwylele
f0ee4c0595 gl_rasterizer: reset texture state context after every draw 2018-06-01 12:05:30 +03:00
James Rowe
7715fd2c19
Merge pull request #3750 from wwylele/cube-watcher-fix
gl_rasterizer_cache: add missing watcher invalidation
2018-05-31 23:11:39 -06:00
James Rowe
f7f5a54bc3
Merge pull request #3751 from wwylele/shader-warning-shutup
gl_shader_gen: rearrange function definition to avoid suprious warnings
2018-05-31 23:10:42 -06:00
James Rowe
e63c374ff0
Merge pull request #3714 from wwylele/primitive-restart-guard
video_core/command_processor: correctly handles 0xFFFF index as a normal index
2018-05-29 23:22:00 -06:00
Markus Wick
caba02d42a gl_rasterizer: Don't flip the texture bindings.
The state object isn't used anywhere else, so there
is no need to revert the state.
And the comment is just wrong: It doesn't matter
which textures are bound on framebuffer binding, it
only matters at draw time. And we reset all bindings
before the draw call. So let's use gl_state as it is
designed to avoid flipping states.
2018-05-28 21:04:59 +02:00
wwylele
874cb42e70 shader/jit: preserve integer & condition register across invocation 2018-05-28 14:41:47 +03:00
wwylele
92a1252835 gl_shader_gen: rearrange function definition to avoid suprious warnings 2018-05-19 00:36:33 +03:00
wwylele
8b4e832c5f gl_rasterizer_cache: add missing watcher invalidation 2018-05-18 23:58:43 +03:00
Markus Wick
8e1e52cad9 gl_rasterizer_cache: Use clean state for glBlitFramebuffer.
Framebuffer blits depends on pixel tests:
Ownership (is fine)
Scissor (is broken on the last commit)
Masking (is broken on master for a while)

So let's be honest and start with a clean state in
those helper functions.
2018-05-18 21:13:56 +02:00
Markus Wick
301073334a gl_rasterizer: Remove redundant scissor state change.
There is no need to disable this state after the draw call,
gl_state will handle this for us if needed. This kind of
redundant state changes are bad for the driver overhead,
as flipping bits will invalidate the driver state.
2018-05-18 21:13:56 +02:00
wwylele
129b893509 gl_stream_buffer: update the information about the AMD hack 2018-05-18 14:08:12 +03:00
wwylele
dd6252a676 gl_rasterizer: fallback to software shader path if buffer overflow happens on hardware shader path 2018-05-18 13:55:19 +03:00
wwylele
6985b13439 [HACK] AMD workaround 2018-05-14 10:17:36 +03:00
wwylele
ede0d15fec video_core/command_processor: attempt accelerate draw in draw trigger 2018-05-14 10:17:36 +03:00
wwylele
9b448a0739 gl_rasterizer: implement AccelerateDrawBatch to emulate PICA shader on hardware 2018-05-14 10:17:36 +03:00
MerryMage
15d14be3cc primitive_assembly: Add getters for internal state 2018-05-14 10:17:35 +03:00
wwylele
06815ec905 video_core: receive hardware shader settings 2018-05-14 10:17:35 +03:00
wwylele
68b0a3e19e regs_pipeline: use proper unsigned type where applicable 2018-05-06 15:57:48 +03:00
Weiyi Wang
f85e71c37c
Merge pull request #3715 from wwylele/hardware-vertex-vector
gl_rasterizer: Use GLvec* instead of C arrays
2018-05-06 07:19:06 +03:00
Weiyi Wang
0da3b75c9e
Merge pull request #3700 from wwylele/texcache-watcher
gl_rasterizer_cache: cache texture cube
2018-05-05 16:30:39 +03:00
Markus Wick
5960282303 gl_rasterizer: Use buffer_storage for uniform data.
This replaces the glBufferData logic with the shared stream buffer code.
The new code doesn't need a temporary staging buffer any more, so the
performance should imrpove quite a bit.
2018-05-05 09:22:02 +02:00