docs(handoff): linux-cpu-profiling — the RAM finding is RETRACTED, and the tree held three efforts - #1136
Merged
Merged
Conversation
…d the tree held three efforts Lands the handoff doc the previous session left in /tmp while its own kickoff block pointed at claudedocs/ — a path that did not exist. Corrected against live state rather than copied. Two substantive corrections: 1. The headline "RAM is 35% below spec, free performance" is retracted. The measurement holds (5600 spec / 3600 actual) but the diagnosis does not. The draft eliminated a per-DIMM cause with "all 4 identical"; inxi -max shows two different part numbers (Crucial Pro CP32G56C46U5 + Crucial CT32G56C46U5) — a mixed kit. Those modules are 2Rx8, so this is 8 ranks on an AM5 controller, which derates to ~3600 by design. CL46 @ 1.1V is the JEDEC 5600 profile, so there is likely no EXPO profile to switch on. 2. The tree held three unrelated efforts, not one. The ranked step "commit nix/pkgs/default.nix + scripts/memory-detail" would have landed half of an unrelated bar feature, without its nix/graphical.nix wiring or tests. All three are now separated onto their own branches. Also records that test_opencode_rig_control.py fails 27/34 under a different $HOME and would turn the required tekton/devrc-pytests check red. Claude-Session-Id: 506ff0fd-a8f2-41c4-a5fc-3b536f342507
…u-profiling-sync Claude-Session-Id: 506ff0fd-a8f2-41c4-a5fc-3b536f342507
…action Its own owner shipped it while this work was in flight. The preservation branch was byte-identical to what landed (verified per file, not by ancestry) and is deleted. Also syncs 63 commits of main so the stale tekton/devrc-pytests result is recomputed against a tree that exists. Claude-Session-Id: 506ff0fd-a8f2-41c4-a5fc-3b536f342507
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Lands the handoff doc the previous session left in
/tmp/opencode/while its own kickoff block pointed atclaudedocs/handoff-linux-cpu-profiling.md— a path that never existed. Corrected against live state rather than copied forward.Retraction. The draft led with "RAM is 35% below rated speed — free performance sitting on the table" and ranked enable EXPO in BIOS as next step #1. The measurement holds; the diagnosis does not. The draft eliminated a per-DIMM cause with "all 4 identical" —
inxi -maxshows two different part numbers (Crucial ProCP32G56C46U5+ CrucialCT32G56C46U5), a mixed kit. Those modules are 2Rx8, so the box runs 8 ranks, which AM5 derates to ~3600 MT/s by design, and CL46 @ 1.1V is the JEDEC 5600 profile — there is likely no EXPO profile to switch on.Three efforts in one tree. The ranked step "commit
nix/pkgs/default.nix+scripts/memory-detail" would have landed half of an unrelated bar feature, without itsnix/graphical.nixwiring or its tests. Separated onto their own branches; see the table in the doc.Also records that
test_opencode_rig_control.py(the file PR #1134's title describes but does not contain) fails 27/34 under a different $HOME and would turn the requiredtekton/devrc-pytestscheck red.Docs only — no code paths touched.