me12345
You can find UTF at pgsdf/UTF. It is a work in progress, and I am implementing semasound (was semaaud) this month so I'll expect to showcase UTF later this year.
Here's information about what it is and its benefits:
UTF provides a unified multimedia substrate for FreeBSD systems. Its main benefit is that graphics, audio, input, and timing all operate on one shared architecture instead of many disconnected subsystems. The project uses an audio driven monotonic clock through chronofs, so sound, rendering, and input events stay synchronized by design. This can reduce timing drift and improve low latency interaction, animation consistency, and audiovisual synchronization.
The system also introduces inputfs, which replaces traditional HID and evdev style handling with a kernel native input layer. UTF parses HID devices directly at the hidbus level and publishes events through shared memory. This design reduces overhead, improves deterministic event ordering, and gives a more unified model for keyboards, mice, touch devices, and gestures. The goal is to simplify compositor and interaction logic while keeping latency low.
Another important benefit is the use of shared memory coordination under /var/run/sema/. UTF avoids many layers of conventional IPC and desktop infrastructure. Processes communicate through shared memory regions for clocks, events, state, and synchronization data. This can reduce serialization overhead and make coordination between graphics, audio, and input services faster and more predictable.
UTF also provides a semantic graphics system called semadraw. Instead of focusing only on low level GPU commands, it represents rendering intent through SDCS streams. This allows the same rendering model to work across multiple backends including EFI framebuffer, Vulkan, X11, Wayland, and software rendering. The architecture aims to improve portability and simplify compositor level optimization.
The platform can operate directly on EFI framebuffers without requiring DRM or modern GPU drivers. This makes the graphics stack simpler and allows graphical operation even on older or unsupported hardware. The project describes this as a way to reduce dependency complexity and maintain a more direct boot to graphics path.
UTF separates the multimedia substrate from the operating system distribution itself. UTF acts as the platform layer, while PGSD defines higher level desktop and distribution choices. This separation makes the architecture more modular and potentially reusable for different environments built on top of the same timing and coordination system.
Overall, UTF is designed around deterministic multimedia coordination. The project treats time as a first class system primitive. By integrating audio timing, rendering, input, and synchronization into one coherent structure, UTF aims to provide lower latency, tighter synchronization, simpler architecture, and more predictable realtime behavior than traditional desktop stacks.
