meta(sciml): expose prepared operators to the SciML integrator stack
还没有人认领这个 Issue。
评估
- 难度
- 5/5
- 预计耗时
- 一周以上
- 新手友好度
- 25/100
- Issue 类型
- 功能
- 描述清晰度
- 基本清楚
- 活跃度
- 活跃
- 技术栈
- julia
调研方向
Treat this as an umbrella rather than a first patch. Read README.md, CLAUDE.md, and DESIGN.md §7, then inspect sub-issues #86, #65, #73, and #87. It is done when the ordered extension work and its constraints have been resolved without adding SciML packages to [deps].
由索引模型根据 Issue 内容生成。
描述
Umbrella for handing MFO operators to OrdinaryDiffEq / LinearSolve. Nothing SciML exists today: no weak dep, no ext/…SciMLExt.jl, and every example hand-rolls its integrator. README.md no longer advertises OrdinaryDiffEq interop that is not there (#87 retracted it; README.md:150 now says there is no integration yet).
Status (2026-09-30): #86 and #87 are done. What remains is #65 first, then #73.
Sub-issues, in order
#86done (PR #89) — prepared-operator trait hygiene —PreparedOperatorhas no supertype andPreparedComposed/PreparedAdjointdeclare noislinear/isconstant, so a wrapper forwarding traits gets thefalsedefault by accident. An invariant hazard independent of SciML; ship first.- #65 — next — the
jac_prototypeadapter (DESIGN §7). Enough thatODEFunction(f; jac_prototype = sciml(prepare(L)))works andKrylovJL_GMREScalls the zero-alloc flatmul!. The downstream forcing function is the microvascular model'sKenCarp47+ sparse-ILU detour. - #73 — after #65 —
SplitODEFunctionfor the linear half plus a per-model work-precision search over IMEX / exponential integrators (a benchmark artifact, not a package feature). #87done (PR #90) — docs — fix the README claim, and document field-levelapply!as the explicit-stepping idiom (the preparedmul!leg pays 15–50% in flat copies that only Krylov needs).
Related, not parented: #63 (stacked multi-species operator) is what the reaction half of a split problem wants to look like; #62 (Robin) is what makes MultigridPreconditioner applicable to the downstream wall-coupled system.
Addendum questions (from the comment below)
- Q1 (what OrdinaryDiffEq requires of a state
u, and the minimal interface that makesFielda legal state) now lives in #73's checklist. - Q3 (ghosts inline in a padded array vs. a flat interior with ghosts beside it in face buffers) is a DESIGN question, not a SciML one, and now lives as a checkbox in #79.
- Q2 (whether a Field-valued Krylov workspace beats the two staging copies per matvec, CPU and GPU separately) is still an open benchmark question here.
Constraints
- Extension only. SciMLOperators / OrdinaryDiffEq are weak deps, nothing enters
[deps](CLAUDE.md, DESIGN.md §7). - Do not route around the
mul!(y, ::AbstractOperator, x::AbstractVector)guard; a convenience constructor must still go throughprepare. - The affine
boundary_rhsterm needs a declared home in a split problem.
—
🤖 Claude tidied this one and moved the addendum questions to their proper homes.
claude-opus-5-5[1m] (high) · claude-code 2.1.285 · unreviewed
- 主要语言
- Julia
- 星标
- 3
- 派生
- 1
- 平均合并
- 4 天 19 小时
- 30 天内合并 PR
- 5
环境准备
这个项目没有提供开发容器、Dockerfile 或贡献指南，环境需要你自己搭建：先看它的 README，通用步骤见我们的新手贡献指南。
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。
RallypointOne/MatrixFreeOperators.jl 的其他 Issue
-
难度 4/5 3-5 天 新手友好度 45/100
-
enhancement
难度 4/5 3-5 天 新手友好度 48/100
-
performance
难度 5/5 一周以上 新手友好度 45/100
-
performance
难度 4/5 3-5 天 新手友好度 48/100
-
enhancement
难度 5/5 一周以上 新手友好度 25/100
查看 RallypointOne/MatrixFreeOperators.jl 的全部 Issue
相似的 Issue
-
mul! with a Symmetric or Hermitian factor throws for a matrix view whose columns are not contiguous未关闭
难度 2/5 1-3 小时 新手友好度 72/100
JuliaLang/LinearAlgebra.jl#1829 ·
维护者通常 1 天内回复
-
难度 2/5 1-3 小时 新手友好度 78/100
MagneticResonanceImaging/MRIReco.jl#320 ·
维护者通常 1 天内回复
-
难度 2/5 1-3 小时 新手友好度 78/100
rfordatascience/tidytuesday#1098 ·
-
难度 2/5 1-3 小时 新手友好度 78/100
ClapeyronThermo/Clapeyron.jl#662 · 2 条评论 ·
-
ci system:windows test
难度 1/5 1 小时以内 新手友好度 76/100
维护者通常 1 天内回复