marfrit
1d0db3b5a9
Measured on hertz (Pi 5, 6.12.75+rpt-rpi-2712, FFmpeg 7.1.3)
to quantify the architectural cost/benefit of routing decode
through the V4L2 m2m + chardev + dmabuf path vs running
ffmpeg standalone.
1080p × 150 frames, decode-as-fast-as-possible:
VP9 8-bit: ffmpeg 214.9% CPU / 1083ms wall
daedalus 96.3% CPU / 1229ms wall
AV1 8-bit: ffmpeg 201.5% CPU / 1162ms wall
daedalus 96.6% CPU / 1478ms wall
H.264 8-bit: ffmpeg 205.8% CPU / 1063ms wall
daedalus 100.1% CPU / 1020ms wall
VP9 10-bit: ffmpeg 155.8% CPU / 269ms wall
daedalus 91.6% CPU / 131ms wall
Key takeaway: the daedalus pipeline uses ~half the CPU for
roughly the same wall throughput. FFmpeg standalone defaults
to 2 threads; for single-stream decode that doesn't
parallelise well, so the 2× CPU usage is overhead, not
parallelism benefit. The daemon's single-threaded serialised
event loop avoids that tax.
For the project's 30fps-floor-is-fine target ("daily YouTube
with CPU free for vscode"), daedalus leaves ~2× the CPU
headroom for the rest of the desktop at the same playback
rate.
VP9-10bit is striking — daedalus is faster wallclock too
(131ms vs 269ms) because at small per-frame work FFmpeg's
thread pool spin-up dominates.
Note: "daedalus" still uses FFmpeg internally (Phase 8.8
explicitly deferred QPU substitution after measurement showed
30fps@1080p was already met). The benefit here is
architectural — single-threaded decode, out-of-process
daemon, dmabuf zero-copy — not QPU offload.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
daedalus-v4l2
V4L2 stateless decoder for the Raspberry Pi 5 / CM5, backed by the
daedalus-fourier kernel library (VP9 + AV1 CDEF + H.264 video
decode kernels on VideoCore VII compute + ARM NEON).
Status: scaffold (2026-05-18). Architecture locked per daedalus-fourier session memory; implementation not yet begun.
What this is
Sibling repo to daedalus-fourier (the kernel library; cycles 1-9 closed).
A two-piece userspace + kernel-module stack that exposes a V4L2
stateless decoder interface (/dev/videoNN) so that
libva-v4l2-request-fourier → firefox-fourier /
chromium-fourier can drive it the same way they drive existing
hardware-decode pipelines on Pi 5 / RK3588.
+-----------------------------------------------------------+
| firefox-fourier / chromium-fourier (existing) |
+-----------------------------------------------------------+
| VA-API |
+-----------------------------------------------------------+
| libva-v4l2-request-fourier (existing, sibling project) |
+-----------------------------------------------------------+
| V4L2 stateless ioctl uAPI |
+-----------------------------------------------------------+
| daedalus-v4l2 kernel module (`kernel/`) |
| - registers /dev/videoNN |
| - parses V4L2 stateless ioctls (VP9/AV1/H.264 controls) |
| - forwards bitstream + controls to userspace daemon |
| via chardev or netlink |
+-----------------------------------------------------------+
| daedalus-v4l2 userspace daemon (`daemon/`) |
| - takes bitstream blobs + per-slice controls |
| - drives FFmpeg parsers via dlopen (Option γ) |
| - dispatches per-block ops via daedalus-fourier |
| public API (daedalus_dispatch_*) |
| - posts decoded frames back to kernel module |
+-----------------------------------------------------------+
| daedalus-fourier kernel library (sibling project) |
| - exports include/daedalus.h public API |
| - per-kernel CPU NEON + opportunistic V3D QPU dispatch |
| - 9 closed cycles across VP9, AV1 CDEF, H.264 |
+-----------------------------------------------------------+
| V3D 7.1 (Mesa userspace v3dv) + ARM NEON (BCM2712) |
+-----------------------------------------------------------+
Why this architecture (Option B + γ + sibling)
Locked by user 2026-05-18 from 3 options in
daedalus-fourier/docs/phase8_scoping.md:
- Option B over A (userspace v4l2loopback): real
/dev/videoNN, proper DRM PRIME / dmabuf for browser zero-copy. - Option γ: dlopen FFmpeg as parser at runtime. No vendoring, fastest to v1.
- Sibling repo: per
project_consumer_targetconvention, V4L2-side work lives outside daedalus-fourier so the kernel-library has a clean API boundary.
Status
Initial scaffold only. See docs/architecture.md for the
deeper design and docs/roadmap.md for the
sub-phase breakdown.
Repo layout
kernel/— Linux kernel module (V4L2 device registration + ioctl handling + userspace chardev bridge). Out-of-tree.daemon/— userspace decoder daemon (linkslibdaedalus_core.afrom sibling daedalus-fourier; uses dlopen for FFmpeg parser).include/— shared headers between kernel and daemon.docs/— architecture + roadmap.
License
Kernel module: GPLv2 (required for kernel-tree compatibility). Userspace daemon: BSD-2-Clause (matches daedalus-fourier).