Scoring & Grades: CPU component is eight subtests, not six Stale since 0.4.0 added the AES-256-GCM and SHA-256 rows. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01SNrEbxEc3Q8FWSU9T5PdCx
Updated loadbearer wiki (markdown)
Updated Home (markdown)
Fix stale "does not measure" claims; note the soak block in compare - Accuracy Notes: the soak test *is* the dedicated sustained-thermal test now, and the memory subtests are no longer all single-threaded — drop those two "not measured" bullets, keep an honest note that soak reads no thermal sensor and runs minutes not hours. Cross-link the thermals section to the soak test. - Comparing Machines: document the SUSTAINED LOAD block that compare shows when every file carries --soak data. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01SNrEbxEc3Q8FWSU9T5PdCx
Note the compare TUI view Comparing-Machines: the interactive scrollable coloured view, its keys, and the --plain / piped fallback. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01SNrEbxEc3Q8FWSU9T5PdCx
Add a wiki sidebar with a command reference _Sidebar.md renders on every page: the page list, a one-line-per-command reference (including the new `score` and `soak`), and a note of what is measured but kept out of the grade. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01SNrEbxEc3Q8FWSU9T5PdCx
Document `loadbearer score` New "Re-scoring without re-running" section on The Baseline covering `loadbearer score` (--baseline / --profile / --curve-k / --output), the before→after banner, and the skip-with-note behaviour for subtests a baseline doesn't cover. Add a Home task-table row and cross-links. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01SNrEbxEc3Q8FWSU9T5PdCx
Note the live soak screen in the TUI Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01SNrEbxEc3Q8FWSU9T5PdCx
Document the soak (sustained-load) test New Benchmark-Methodology section covering the soak kernel, the per-second sampling, and the peak / steady / retained / onset / stability figures. Note it in the Home task table, the Scoring page (ungraded, like the link probe), and the result-file field table on The Baseline. Also update the all-core memory-read row for the 0.4.x barrier + 32 MiB floor change. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01SNrEbxEc3Q8FWSU9T5PdCx
Document the AES/SHA and all-core-read subtests
Document single-threaded core pinning
Network is no longer folded into the overall grade
Document the network component and the --net-target link test
Populate the wiki: methodology, scoring, compare, baseline, config, accuracy Home page indexes six deep-dive pages that expand on what the README summarises, cross-linking back to it rather than duplicating it.
Initial Home page