Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

WIP: Metal backend #1287

Open
wants to merge 188 commits into
base: main
Choose a base branch
from
Open

WIP: Metal backend #1287

wants to merge 188 commits into from
+12,463 −296

Conversation

SamoZ256
Copy link

@SamoZ256 SamoZ256 commented Aug 12, 2024 •
edited
Loading

This PR implements a Metal backend for macOS. It should bring better performance compared to Vulkan, as well as fix some issues that wouldn't be possible with Vulkan due to MoltenVK limitations. I haven't tested many games, but those that I tried worked pretty well.

TODO:

  • sRGB
  • fix low resolution
  • proper present
  • fix black textures in sm3dw
  • buffer copies in a vertex shader
  • pipeline cache with binary archives
  • geometry shaders
  • GamePad screen in a separate window
  • fix texture readback
  • output shader
  • ImGui
  • Intel Macs support
  • proper surface copy
  • extend the shader hash
  • triangle fans
  • 1D shadow sampler emulation
  • add an option to turn mesh shaders off
  • instancing with mesh shaders
  • flush tile memory using tile shaders
  • occlusion queries

Screenshots:

New Super Mario Bros U nsmbu_missing_geometry_fixed
Captain Toad Treasure Tracker captain_toad_lvl2_working
Super Mario 3D World sm3dw_water_fixed
Mario Kart 8 mk8_black_artifacts_fixed
The Legend of Zelda: Breath of the Wild botw
Super Mario Maker smm_missing_sprites_fixed

SamoZ256 and others added 30 commits July 25, 2024 11:18
@SamoZ256
start metal backend
2477bad
@SamoZ256
initialize Metal
9837026
@SamoZ256
implement texture view & fix: crashes
6db893c
@SamoZ256
create and present drawable
9b127be
@SamoZ256
implement pixel formats and texture copy
46981d7
@SamoZ256
start the shader decompiler
6959716
@SamoZ256
improve shadergen & create shaders
aac9b12
@SamoZ256
fix: shader errors and shader crashes
af3ce80
@SamoZ256
fix: sampling & capture boundaries
4976ff3
@SamoZ256
implement texture decoding
4c78c6a
@SamoZ256
implement render passes
e539527
@SamoZ256
implement vertex descriptors & draw
cb525b2
@SamoZ256
bind resources to shaders
7ea18d8
@SamoZ256
bind default sampler
35eea12
@SamoZ256
present
7ae29a7
@SamoZ256
update CMakeLists.txt
8800de0
@SamoZ256
Merge branch 'cemu-project:main' into metal
89a2c23
@SamoZ256
rework command encoder system
f011300
@SamoZ256
fix: support buffer data & fix: depth
0cb83d4
@SamoZ256
optimize index buffers
be8a560
@SamoZ256
set viewport and scissor
1fbd6ad
@SamoZ256
fix: viewport and scissor
d4564c1
@SamoZ256
implement blending
e4abb30
@SamoZ256
fix: shader errors
a2d1585
@SamoZ256
implement depth state & texture clears
4173675
@SamoZ256
implement proper bit cast in shaders
4022755
@SamoZ256
fix: present issue
db709c3
@SamoZ256
fix: viewport
c6f6616
@SamoZ256
implement render pipeline cache
d7411e2
@SamoZ256
implement depth stencil cache & fix: textureSize
406a856
SamoZ256 and others added 30 commits September 4, 2024 19:05
@SamoZ256
don't compile fragment shaders just-in-time
8a74445
@SamoZ256
include color format data types in ps hash
cd72ad8
@SamoZ256
check for invalid color formats
4251f3f
@SamoZ256
refactor pixel format support
6a3bdd4
@SamoZ256
add: todo notices
548ffb6
@SamoZ256
Merge pull request #3 from SamoZ256/metal-shaders
5e50592 
Better shaders
@SamoZ256
support rasterization kill
3dc233f
@SamoZ256
optimize rasterization kill
cf56024
@SamoZ256
make binary archives device and os version specific
8404981
@SamoZ256
use lcr
e8f726e
@SamoZ256
handle rasterization kill for mesh shaders
395cd1c
@SamoZ256
support instancing for mesh shaders
950f04d
@SamoZ256
simplify uniform names
e7f8f0e
@SamoZ256
use uchar for index type
a328c5e
@SamoZ256
put query object into a separate file
4cce369
@SamoZ256
implement occlusion queries
eb7c10e
@SamoZ256
improve command buffer sync
fd16488
@SamoZ256
handle occlusion queries with no draws
934b1f8
@SamoZ256
request soon commit instead of committing directly
e89efed
@SamoZ256
Merge pull request #4 from SamoZ256/metal-occlusion-queries
e5dcd93 
Occlusion queries
@SamoZ256
release command buffers properly
b5954d8
@SamoZ256
Merge branch 'cemu-project:main' into metal
25a1d7e
@SamoZ256
set pixel format view usage for textures
02254d4
@SamoZ256
allocate special buffers only when needed
008c11c
@SamoZ256
correct comment
2961151
@SamoZ256
make a workaround for streamout with no fbo
358567a
@SamoZ256
use correct stages for buffer sync
8ac90ce
@SamoZ256
fix: invalid surface copy depth
535107e
@SamoZ256
refactor pixel formats
aa81070
@SamoZ256
only set buffer offset if needed
f03c0a2
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers
No reviews
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
3 participants
@SamoZ256 @Exzap @Spritzerland