ca801b6445
the HunkAlloc size calculated in calTexinfoAndFacesSize() was too
small because that map has a *massive* water poly.
Unfortunately it's not feasible to calculate the correct size
(it depends on the size of the poly, for which I'd need the vertex
coordinates, but to get them I'd have to load them first, which requires
the allocation...), but allocating 5MB extra (instead of 1) made it work
|
||
---|---|---|
shaders | ||
src | ||
.gitignore | ||
LICENSE | ||
Makefile | ||
README.md |
Vulkan Renderer Library for Yamagi Quake II
This is the vkQuake2 vulkan renderer library ported to Yamagi Quake II.
Compilation
You'll need:
- clang or gcc,
- GNU Make,
- SDL2 with
sdl2-config
, - vulkan-headers,
- vulkan-validationlayers if you like to debug issues.
Type make
to compile the library. If the compilation is successfull,
the library can be found under release/ref_vk.dll (Windows) or
release/ref_vk.so (unixoid systems).
Usage
Copy the library next to your Yamagi Quake II executable. You can select
Vulkan through the video menu or by cvar with vid_renderer vk
followed
by a vid_restart
.
If you have run into issues, please attach output logs with OS/driver version and device name to the bug report. List of currently tested devices for the reference.
Note: Game saves outputs to Documents\YamagiQ2\stdout.txt
under
windows.
Console Variables
-
r_validation: Toggle validation layers:
0
- disabled (default in Release)1
- only errors and warnings, show image load issues2
- best-practices validation
-
vk_strings: Print some basic Vulkan/GPU information.
-
vk_mem: Print dynamic vertex/index/uniform/triangle fan buffer memory usage statistics.
-
vk_device: Specify index of the preferred Vulkan device on systems with multiple GPUs:
-1
- prefer first DISCRETE_GPU (default)0..n
- use device #n (full list of devices is returned byvk_strings
command)
-
vk_sampleshading: Toggle sample shading for MSAA. (default:
1
) -
vk_flashblend: Toggle the blending of lights onto the environment. (default:
0
) -
r_polyblend: Blend fullscreen effects: blood, powerups etc. (default:
1
) -
vk_skymip: Toggle the usage of mipmap information for the sky graphics. (default:
0
) -
vk_finish: Inserts a
vkDeviceWaitIdle()
call on frame render start (default:0
). Don't use this, it's there just for the sake of having agl_finish
equivalent! -
vk_custom_particles: Toggle particles type:
0
- textured triangles for particle rendering1
- between using POINT_LIST (default)2
- textured square for particle rendering
-
vk_particle_size: Rendered particle size. (default:
40
) -
vk_particle_att_a: Intensity of the particle A attribute. (default:
0.01
) -
vk_particle_att_b: Intensity of the particle B attribute. (default:
0
) -
vk_particle_att_c: Intensity of the particle C attribute. (default:
0.01
) -
vk_particle_min_size: The minimum size of a rendered particle. (default:
2
) -
vk_particle_max_size: The maximum size of a rendered particle. (default:
40
) -
vk_picmip: Shrink factor for the textures. (default:
0
) -
vk_pixel_size: Pixel size when rendering the world, used to simulate lower screen resolutions. The value represents the length, in pixels, of the side of each pixel block. For example, with size 2 pixels are 2x2 squares, and at 1600x1200 the image is effectively an upscaled 800x600 image. (default:
1
) -
vk_dynamic: Use dynamic lighting. (default:
1
) -
vk_showtris: Display mesh triangles. (default:
0
) -
vk_lightmap: Display lightmaps. (default:
0
) -
vk_postprocess: Toggle additional color/gamma correction. (default:
1
) -
vk_mip_nearfilter: Use nearest-neighbor filtering for mipmaps. (default:
0
) -
vk_texturemode: Change current texture filtering mode:
VK_NEAREST
- nearest-neighbor interpolation, no mipmapsVK_LINEAR
- linear interpolation, no mipmapsVK_MIPMAP_NEAREST
- nearest-neighbor interpolation with mipmapsVK_MIPMAP_LINEAR
- linear interpolation with mipmaps (default)
-
vk_lmaptexturemode: Same as
vk_texturemode
but applied to lightmap textures. -
vk_underwater: Warp the scene if underwater. Set to
0
to disable the effect. Defaults to1
.
Console Variables (MacOS)
-
vk_molten_metalbuffers: enable/disable Metal buffers to bind textures more efficiently (>= Big Sur). (default:
0
) -
vk_molten_fastmath: enable/disable float point op optimisations. (default:
0
)
Custom model format support
Render unofficially supports mdl/Quake 1, dkm/Daikatana and fm/Heretic2, are provided without any warranty of support. The simplest way to check is renaming the mdl/dkm/fm format file to md2 and place instead the original tris.md2 file. FM is rendered with all meshes without support of filtering/selecting the exact part of the model.