Menu
The review-first prompt is selected by default. You can switch to a direct command or download a local copy.
Read SKILL.md and any companion files shown by SkillsMP before deciding whether to install.
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
A direct command skips the review prompt. Inspect the source before running it.
npx skills add https://github.com/ZhangXin8069/20250814-20260815 --skill snsc-v20260726The command stays on one line. Scroll horizontally to inspect it before copying.
Prefer a local copy? Download the files currently available to SkillsMP.
Based on SOC occupation classification
Showing SKILL.md
| name | snsc-v20260726 |
| description | 运行 snsc-v20260726 CPU 集群管线 — 剥离 AI agent 阶段的 NumPy 验证管线 (已被 GPU 管线取代) |
在 /root/lattice-pdf/agent/snsc-v20260726 运行剥离了 AI agent 阶段（LQCD_Master + lamet-agent）的 CPU (NumPy) disconnected 胶子 PDF 验证管线。
⚠️ 已被 GPU 管线 ../docker-v20260802/（及更新版本）取代。 无 AI agent，纯物理计算。
系综: beta6.20_mu-0.2770_ms-0.2400_L24x72
格点: 72×24³, a=0.1053 fm, β=6.20
Nev: 100
组态: 6250, 6450, 6650
动量: P = (0, 0, -2)
算符: Cγ₅γ₄
cd /root/lattice-pdf/agent/snsc-v20260726
# 直接运行
python run_pipeline.py
# 选择性步骤
python run_pipeline.py --skip-2pt --skip-ope --skip-analysis --skip-report
# Slurm 集群提交
sbatch sbatch.sh
# 先从集群下载数据
bash download_data.sh --yes --skip-existing
0 环境检查 → 1 质子 2pt 蒸馏 (CPU) → 2 OPE 从头计算或加载预计算 donghx 数据
→ 3 huangcl 比率分析 (R(z) = C3_disc/C2, jackknife) → 4 Markdown 报告
| 文件 | 用途 |
|---|---|
run_pipeline.py | 4 步主编排器 (计时/内存跟踪) |
compute_2pt.py | 质子 2pt 蒸馏 (CPU) |
compute_ope.py | OPE 算符 (ILDG .lime → Clover F_{μν}) |
analyze_ratio.py | huangcl 比率分析 + jackknife |
gamma_matrix.py / utils.py | γ 矩阵 / 共享工具 |
download_data.sh / sbatch.sh | 数据下载 / Slurm 提交 |
依赖 ../../snsc/main.py 的 plaquette 构造。数据从 SNSC 集群经 download_data.sh (rsync/scp) 下载。