jessey
released this
2026-05-29 15:01:01 +00:00
| 0 commits to main since this release
Added
GPU stress test + close thermal monitoring (rigdoctor stress, and a "Stress test…" button
on System Health). Runs a GPU load and samples sensors at a high rate (default 0.5 s), then
reports per-metric min/avg/peak, how long the core spent above each temperature threshold,
power vs the limit, throttling (decoded from the NVML clocks-event bitmask), and any GPU fault
(Xid / VA-space freeze / a query-timeout hang) that hit during the window — the on-demand way to
reproduce load-correlated crashes. The load comes from an explicit --command (a game or a tool
like gpu-burn), an auto-detected loader (gpu-burn/vkmark/glmark2/vkcube), or monitor-only when
none is found (you launch the game; it tracks temps while you play).
Drive health & wear in the health report. A new core/drives.py parses the full smartctl --json for every drive into prioritized findings: the SMART verdict, a derived life-left % (NVMe percentage_used or the SATA wear-leveling attribute), power-on hours,
data written (TBW), temperature, and the early-failure predictors (reallocated / pending /
offline-uncorrectable sectors, NVMe media errors, low available spare). Replaces the old
pass/fail-only SMART check; flows through the same elevated path (GUI launch / sudo rigdoctor report), degrading to per-drive "needs root" notes unprivileged.
Fixed
GUI "Add game…" can now link a launcher. The dialog only asked for a name, so a custom
game (e.g. SPT) couldn't be given its launch command or log folder from the app — those were
CLI-only, leaving it unlaunchable from the GUI. It's now a proper form: name + an optional
launch command/script (with a Browse… file picker) + an optional log folder (auto-detected
from the script's folder when left blank).
jessey released this