[Paper] Measure, then clean up the `_s` problem variants
维护者通常 1 天内回复
@ocots 已经在做这个了。
开始于 2026年10月31日。
评估
这个 Issue 还没有评估数据。
描述
Goal
Decide whether the _s problem variants (dynamics written coordinate by coordinate, historically required by Exa) are still useful. Decision taken: measure first whether the scalarized version is significantly better than the vectorial one. If it changes nothing, remove the _s variants; otherwise keep both and compare them in the paper.
What we know
_svariants exist for the 22 problems (ext/OptimalControlModels_s/), with their stubs andtest/test_OptimalControl_s.jl.- In OptimalControl v2,
:exaaccepts the vector form of the dynamics (see the GPU page of the documentation), so the original reason may be gone. CTBenchmarks still runs Exa through<problem>_s. - The coordinate-wise form changes the ExaModels code generated by CTParser, hence possibly the number of expression nodes and the GPU performance.
Proposed measurements
- 22 problems,
:exa(CPU and GPU) and:adnlp; three grid sizes (for instance 500, 2000, 10000). - Model build time, solve time, iterations, objective, nonzeros of the Jacobian and of the Hessian.
- Criterion to fix (for instance median time difference below about 10 %, same iterations and objective).
If the variants are removed
Delete ext/OptimalControlModels_s/, the <problem>_s stubs and the corresponding test; close #174 and #136.
Related
#174, #136. Depends on the v2 port of both packages for a clean measurement.
How to work on this theme
This is a theme issue: it describes the work to do and what is known so far, it is not meant to be closed by a single PR.
- Whoever takes it first defines the work precisely and splits it into sub-issues (they can live in any repository of the organization).
- Work proceeds through the sub-issues; tick the checklist below as they are closed.
- Close the theme when everything is done or explicitly dropped.
Context: tracking issue control-toolbox/CTBenchmarks.jl#200; planning notes: objectives and directions and updating OptimalControlProblems and CTBenchmarks (in control-toolbox/benchmark-paper).
Sub-issues
- to be created when the theme is taken
- 主要语言
- Julia
- 星标
- 13
- 派生
- 1
- 平均合并
- 3 小时 24 分钟
- 30 天内合并 PR
- 34
环境准备
- 没有 Dockerfile 或 Docker Compose 文件
- 没有 Pull Request 模板
- 阅读贡献指南
从这里开始
- 先读完整个 Issue，再读项目的贡献指南。
- 在 Issue 下留言说明你要接手 —— 这能避免两个人做同样的事。
- Fork 仓库，在一个分支上完成修改。
- 提交 Pull Request，并在描述里引用这个 Issue 编号。
control-toolbox/OptimalControlProblems.jl 的其他 Issue
-
[Dev] One source of truth per problem, and the equality of the modelings checked by computation可能已有人在做 @ocots 今天认领。 未关闭
control-toolbox/OptimalControlProblems.jl#391 · 已指派 1 人 ·
维护者通常 1 天内回复
-
cron-failure
难度 4/5 1-2 天 新手友好度 25/100
control-toolbox/OptimalControlProblems.jl#381 ·
维护者通常 1 天内回复
-
[Metadata] Declared fields: tags, references, Corleone, reference solution可能已有人在做 @ocots 于 1 天前认领。 未关闭
control-toolbox/OptimalControlProblems.jl#379 · 已指派 1 人 ·
维护者通常 1 天内回复
-
[Metadata] Shared initial guess in the metadata, read by every modeling可能已有人在做 @ocots 于 1 天前认领。 未关闭
control-toolbox/OptimalControlProblems.jl#377 · 1 条评论 · 已指派 1 人 ·
维护者通常 1 天内回复
-
[Dev] Precompile the problems so that users and tests do not recompile them every time可能已有人在做 @ocots 于 1 天前认领。 未关闭
control-toolbox/OptimalControlProblems.jl#373 · 1 条评论 · 已指派 1 人 ·
维护者通常 1 天内回复
查看 control-toolbox/OptimalControlProblems.jl 的全部 Issue
相似的 Issue
-
难度 2/5 1-3 小时 新手友好度 68/100
维护者通常 1 天内回复
-
Add project: Typst_jll.jl可能已有人在做 @github-actions 于 1 天前认领。 未关闭add-project
难度 1/5 1 小时以内 新手友好度 76/100
YDX-2147483647/best-of-typst#590 · 1 条评论 ·
维护者通常 1 天内回复
-
难度 2/5 1-3 小时 新手友好度 74/100
维护者通常 1 天内回复
-
难度 2/5 1-3 小时 新手友好度 74/100
-
难度 2/5 1-3 小时 新手友好度 67/100