Fix per-game settings (Issue #115)
Automatic load/save state (Issue #173)
Reload Naomi/AW BIOS after per-game settings are loaded to allow
per-game BIOS region. (Issue #96)
Restrict Soul Reaver widescreen cheat to the EU version (Issue #188)
Disable UI elements for overridden settings
Clean up/simplify legacy Naomi ROM loading
Increase descriptor set pool size for combined image sampler to
accomodate for games using a stupidly large number of textures.
Don't use mailbox mode on AMD as it needs 5(!) swap chain images on
linux.
Return to a swap chain size of 3 since it's the min for AMD.
move ui loop out of hw/pvr
vk: implement multi-rendering, delay swap. Load fb before render pass
fps counter now counts frames, not renders
vk: force ignoreTexAlpha for 565 texture in case the supporting texture
has an alpha channel (RTT). Fixes Alone in the Dark cinematic scenes.
pvr: better detection of frame swapping (FB_R_SOF/FB_W_SOF)
pvr: synchronous rendering now always waits for render in progress
(speed measurement not reliable, render skipped when multi-rendering)
vk: limit swap chain size to 2
linux: look for legacy ~/.reicast and ~/.reicast/data
else look for ~/.config/flycast and ~/.local/share/flycast
and ~/.config/reicast and ~/.local/share/reicast
(defaults to flycast)
look for bios files in home folder and data folders (android, windows
,macos), then in game folder.
on linux, search in /usr/share/flycast and /usr/local/share/flycast and
legacy locations
Linear tiling is faster for small, frequently updated textures
When lots of textures are updated each frame, a skipped frame will
destroy/recreate all textures, causing another skipped frame, etc. So
in-flight texture tracking is disabled, except for RTT textures
Limit swap chain to 2 images
Refactor base and oit renderers to avoid code dup
BufferData was always allocated in host visible mem, regardless of the
mem property argument.
NVidia gpus don't support blitting to 565 and 5551, which is needed for
mipmaps, so convert to 8888 instead.
Use optimal tiling for oit pointers image (linear not supported by
nvidia and probably slower anyway)
oit screen drawer uses multiple final attachments
Set non-oit depth attachment as transient and lazily allocate mem
Flyinghead