Glassvision/JuliaFEM.jl
1
0
Fork 0
mirror of https://github.com/JuliaFEM/JuliaFEM.jl.git synced 2026-08-06 04:21:33 +00:00
Code Issues Packages Projects Releases Wiki Activity

docs(agents): link zero-alloc invariant to performance tiers

Browse Source 
Point invariant 1 readers at `architecture_layers.md` for the tiered split
between numeric kernels, warmed assembly drivers, and flexible IO/UI code.
This commit is contained in:
Jukka Aho
2026-05-09 18:23:24 +03:00
parent f0486eae35
commit 83cd1a4407
1 changed files with 5 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files
AGENTS.md
+5 -1
View File
@@ -176,7 +176,11 @@ These are non-negotiable. Tests and code analysis enforce them.
`test/assemblers/test_dof_based_zero_alloc.jl` asserts
`@allocated assemble!(...) == 0` and `0` GC allocation sites in the
optimized LLVM IR. Don't introduce `Dict`, `Vector{Any}`, untyped
closures, or `Vector` literals inside loops.
closures, or `Vector` literals inside loops. For the intended split
between tier 1 numeric kernels (always C-speed, no heap churn in
loops), tier 2 warmed assembly drivers (setup may allocate), and tier 3
IO/UI convenience code where flexible containers are acceptable, see
`docs/src/developer/architecture_layers.md` (section Performance tiers).
2. Type stability everywhere on the hot path.
`Base.promote_op(assemble!, ...) === Nothing` and the inferred
types must be concrete. Use `NamedTuple` (typed), `NTuple`, and