Adds an experimental power-cap mode using the undocumented RM ioctl
interface (based on panchovix's LACT PR #1205) to set power limits
below the VBIOS minimum (down to 30 W).
- hal/rm_power.py: RM ioctl power-cap read/write/reset + runtime probe
- limits.py: power_cap_mode (nvml/ioctl) with support detection
- config.py: persist power_cap_mode per GPU
- profiles: record/apply power_cap_mode
- server.py: POST /api/limits validates ioctl support (409 on failure)
- cli.py: profile save falls back to persisted mode
- client.py: power_cap_mode in Limits
- frontend: toggle + warning with panchovix attribution (LACT #1205)
- tests: test_rm_power.py (unit) + integration coverage
- Makefile: add test_rm_power.py to make test
Also includes automated linter reformatting (prettier, ruff, shellcheck,
isort, markdownlint) that the linter would apply anyway.
- hatch_build.py: custom hatchling build hook that compiles the React
frontend (npm ci + build) when frontend/dist is missing or stale, so
'uv tool install git+https://gitea.zephyre.one/Pakobbix/nvcurve.git'
works as a single command
- install.sh: curl|bash installer (checks prerequisites, auto-installs uv,
clones and installs)
- Makefile: dev targets (frontend, install, dev, test, clean)
- README/docs: document the one-liner, clone, and direct-git installs
- Replace PyPI-only install with full source build workflow (npm + uv)
- Add Blackwell GPU fork notice with extended memory offset and fan control
- Add systemd service section: install commands, unit file template, config reference
- Update docs/Installation.md to match npm + uv tool install . workflow
- Update docs/Overview.md with Blackwell-specific capabilities
ARIA