flow_solver_staggered.cpp File Reference

The ONE compiled instantiation of the staggered MAC solver, peclet::flow::Solver<Staggered>. More...

#include "anderson_accelerator.hpp"
#include "flow_ibm.hpp"

Go to the source code of this file.

Namespaces
namespace	peclet
namespace	peclet::flow

Macros
#define	PECLET_FLOW_INSTANTIATING   1

Detailed Description

The ONE compiled instantiation of the staggered MAC solver, peclet::flow::Solver<Staggered>.

Solver<Grid> is a ~12 k-line class template whose 517 out-of-line members live in the twelve flow_ibm_*.hpp domain headers. Before QUALITY_PLAN G.8 every consumer — 45 test executables and the nanobind module, the latter for both grids — instantiated all of it from scratch at -O3, which is what made a full rebuild ~45-50 CPU-minutes. This translation unit instantiates it once; flow_ibm.hpp carries the matching extern template declaration, which suppresses the implicit instantiation everywhere else (PECLET_FLOW_INSTANTIATING skips it here).

There is NO cross-TU optimization to lose: the build carries no LTO and no -march, so the kernels compiled here are the same kernels, from the same source, under the same flags, that each consumer used to compile for itself. Members defined inside the class body stay implicitly inline and are therefore still instantiated (and inlined) in the consumer — the standard exempts inline functions from an explicit instantiation declaration — so the small accessors on the hot path are unaffected.

PECLET_FLOW_MPI must agree between this TU and its consumers: it changes the class definition. The build therefore compiles this file once per configuration (peclet_flow_solver without the macro, peclet_flow_solver_mpi with it) and links each consumer against the matching one.

Definition in file flow_solver_staggered.cpp.

Macro Definition Documentation

PECLET_FLOW_INSTANTIATING

#define PECLET_FLOW_INSTANTIATING   1

Definition at line 22 of file flow_solver_staggered.cpp.