Repository navigation
docs: 补记缓存的实测收益(~19s/17%),并纠正我先前错误的前提判断 - #11
Merged
Merged
Conversation
实测(三次运行对比): PR #10(下载,miss→save) 116s 主干 push(下载,miss→save) 99s 主干 dispatch(**命中**) 97s ⇒ 缓存实际只省 ~19s(~17%)。 **推翻我原先的说法**:本 DEC 开头写「下载占大头」——**是错的**。 GitHub runner 网络很快,1.2 GB 约 20 秒下完;那 ~100 秒里主要是 Ollama 二进制安装 + pip + 测试本身(CPU 推理)。正确说法是「下载约占 1/5」。 来源分析:原判断来自**本机**体感(8 GB 内存 + 慢网络),**不能外推到 CI** —— 与 DEC-005(把当次测量当成架构事实)是同一族错误。 保留该缓存的理由(已写进 DEC-009 §五之三): · 仍是净收益;成本只有 998 MB 配额(上限远未用满) · 为正确性加的三件事(停 systemd 服务 · 钉死 OLLAMA_MODELS · 守卫)与缓存无关也该有 —— 否则模型落在哪不确定,测试跑的可能不是你以为的那份权重 判据:**收益小 ≠ 该撤销,但必须把真实收益说清楚** —— 连"省多少"都不量,就会一直拿一个夸大的理由维护一段复杂度。 提交前两问 ① 本次无新增决策事项(沿用 DEC-009,只是补记实测数据与纠正前提) ② CHANGELOG 已写 测试:61 passed
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
实测数据(三次运行对比)
pytest(含 embedding)⇒ 只省 ~19 s(~17%)。
我在
DEC-009开头写的是「下载占大头」—— 是错的。原因:GitHub runner 的网络很快,1.2 GB 大约 20 秒下完;那 ~100 秒里主要是
Ollama 二进制安装 + pip install + 测试本身(CPU 推理)。正确说法是「下载约占 1/5」。
为什么我会判断错:那个量级感来自本机(8 GB 内存 + 慢网络),不能外推到 CI —— 与本仓
DEC-005(把当次测量当架构事实)是同一族错误。那还留着它吗?—— 留
OLLAMA_MODELS· 守卫)与缓存无关也该有 ——否则模型落在哪个目录是不确定的,测试跑的可能不是你以为的那份权重
提交前两问
① 无新增决策事项(沿用
DEC-009,只补实测数据 + 纠正前提)② CHANGELOG 已写