Issues

• Device LDLᵀ: serial pivot search, one workgroup per regime-C front, no vendor sytrf (T25 follow-up from T15)

  found-by-agenttriaged
  exanauts/SparseDirectSolver.jl#75 · Filter by author claude opened on Oct 2, 2026
  1

• K2 dumps: max|L| 1e14–1e16 and factor error up to 3.6 with every pivot_pairs mode; needs scaling (T21) or delayed pivots (M13)

  found-by-agenttriaged
  exanauts/SparseDirectSolver.jl#71 · Filter by author michel2323 opened on Oct 2, 2026
  1

• 2×2 pivot pairs: fixed candidate tolerance τ = 1e-6 and greedy matching; matching-based pairs (T21)

  found-by-agenttriaged
  exanauts/SparseDirectSolver.jl#67 · Filter by author michel2323 opened on Oct 2, 2026

• Regime-A follow-ups: owed CUDA timings and per-backend local-memory caps

  found-by-agenttriaged
  exanauts/SparseDirectSolver.jl#60 · Filter by author claude opened on Oct 2, 2026

• Dense :ka/:generic paths and the KA CPU backend allocate in the numeric phase (Julia 1.10, coverage)

  found-by-agenttriaged
  exanauts/SparseDirectSolver.jl#53 · Filter by author claude opened on Oct 2, 2026
  2

• External — MadNLPGPU integration (in the MadNLP repository)

  externaltask
  exanauts/SparseDirectSolver.jl#28 · Filter by author michel2323 opened on Oct 1, 2026 · M4 — Symmetric indefinite LDLᵀ/LDLᴴ (MadNLP-ready)

• T27 — Robustness extras

  task
  exanauts/SparseDirectSolver.jl#27 · Filter by author michel2323 opened on Oct 1, 2026 · M13 — Robustness extras

• T26 — Hybrid memory / hybrid execute

  task
  exanauts/SparseDirectSolver.jl#26 · Filter by author michel2323 opened on Oct 1, 2026 · M12 — Hybrid modes

• T25 — Performance pass

  task
  exanauts/SparseDirectSolver.jl#25 · Filter by author michel2323 opened on Oct 1, 2026 · M11 — Performance

• T24 — ND partition-tree export/import and ordering cache

  task
  exanauts/SparseDirectSolver.jl#24 · Filter by author michel2323 opened on Oct 1, 2026 · M1 — Symbolic engine

• T23 — AMDGPU, oneAPI and Metal extensions

  task
  exanauts/SparseDirectSolver.jl#23 · Filter by author michel2323 opened on Oct 1, 2026 · M2 — Numeric kernels

• T22 — Non-uniform batch (BatchedDirectSolver)

  task
  exanauts/SparseDirectSolver.jl#22 · Filter by author michel2323 opened on Oct 1, 2026 · M10 — Non-uniform batch

• T21 — Matching and scaling

  task
  exanauts/SparseDirectSolver.jl#21 · Filter by author michel2323 opened on Oct 1, 2026 · M9 — Matching and scaling

• T20 — Schur complement mode

  task
  exanauts/SparseDirectSolver.jl#20 · Filter by author michel2323 opened on Oct 1, 2026 · M8 — Schur complement mode

• T19 — General LU ("G"): CPU reference + GPU

  task
  exanauts/SparseDirectSolver.jl#19 · Filter by author michel2323 opened on Oct 1, 2026 · M7 — General LU (LDU)

• T18 — FGMRES-IR extension (Krylov.jl)

  task
  exanauts/SparseDirectSolver.jl#18 · Filter by author michel2323 opened on Oct 1, 2026 · M5 — Solve extras

• T17 — Uniform batch (v1 cut line)

  claude:implementtask
  exanauts/SparseDirectSolver.jl#17 · Filter by author michel2323 opened on Oct 1, 2026 · M6 — Uniform batch (v1 cut line)
  2