Skip to content

Add per-trial CPU and scheduling evidence to model timing - #38

Merged
PyYoshi merged 1 commit into
devfrom
v3/timing-resource-observation
Sep 20, 2026
Merged

PyYoshi merged 1 commit into
devfrom
v3/timing-resource-observation

Conversation

@PyYoshi

@PyYoshi PyYoshi commented Sep 20, 2026

Copy link
Copy Markdown
Owner

概要

  • Linuxの単一thread診断processでwall計測前後のRUSAGE_SELF差分を取得
  • user/system CPU時間、voluntary/involuntary context switch、minor/major faultを全試行で保存
  • report schema v2。resourceとelapsed配列は同じ順序で保持し、遅い試行を除外しない
  • CPU時間はmicrosecond精度でありns分解能ではないこと、wall/resource区間差と原因推定の限界を明記
  • engine・table・公開APIは不変。non-Linuxは未観測null、Linux runnerはresource不足を拒否

検証

  • native環境付きモデルsuite 87 tests成功
  • Clang sequence probe 10 tests成功
  • 3 run × 16文書 × 3モデル × 7試行で全snapshot/checksum一致、resource配列の件数/順序を照合
  • Ruff E/F、git diff --check

新しい2件のtrial変動ではCPU時間も増加。OS待ちだけと断定せず、未解決として保持。詳細な結果はcChardet側で文書化する。P01は再開しない。

@PyYoshi
PyYoshi marked this pull request as ready for review September 20, 2026 22:48
@PyYoshi
PyYoshi merged commit 10ee109 into dev Sep 20, 2026
11 checks passed
@PyYoshi
PyYoshi deleted the v3/timing-resource-observation branch September 20, 2026 22:48
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant