|
waLBerla 7.3
|
Namespaces | |
| namespace | detail |
| namespace | exectag |
| namespace | internal |
Classes | |
| class | BorderSweep |
| Sweep restriction to a domain border. More... | |
| class | FusedSweep |
| Fusion of two sweeps. More... | |
| class | CommHidingSweep |
| class | SweepFactory |
| Factory for derived and wrapper sweeps. More... | |
| struct | CellIdx |
| class | SparseIndexList |
| class | SerialSweeper |
Enumerations | |
| enum | CommHidingIntervall { ALL , INNER , OUTER } |
Functions | |
| template<CellFunction Func> | |
| void | forAllCells (const exectag::Serial &, const CellInterval &ci, Func &&func) |
| Single-thread CPU cellwise kernel dispatcher. | |
| template<CellFunction Func> | |
| void | forAllCells (const exectag::OpenMP &, const CellInterval &ci, Func &&func) |
| Multi-threaded CPU cellwise kernel dispatcher using OpenMP. | |
| template<CellFunction Func> | |
| void | forAllCells (const exectag::GPU &, const CellInterval &, Func &&) |
| GPU-accelerated kernel dispatcher. | |
| template<ExecTag XTag, CellFunction Func> | |
| void | forAllCells (const XTag &exTag, const StructuredBlockForest &blocks, Func &&func) |
| Invoke forAllCells on the default per-block cell interval. | |
| void | sync () |
| Synchronize with concurrent execution on the default stream. | |
| template<ExecTag XTag, CellPairFunction Func> | |
| void | forAllCellPairs (XTag xtag, const CellInterval &firstCi, const CellInterval &secondCi, Func func) |
| Run an operation for every pair of cells in the given intervals. | |
|
inline |
Run an operation for every pair of cells in the given intervals.
Iterates the given cell intervals sequentially and in lockstep in z-y-x order (x is the fastest, z the slowest coordinate) and calls func for every pair of cells.
| std::invalid_argument | In debug mode, throws if the given cell intervals are of different size. |
| void walberla::v8::sweep::forAllCells | ( | const XTag & | exTag, |
| const StructuredBlockForest & | blocks, | ||
| Func && | func ) |
Invoke forAllCells on the default per-block cell interval.
|
inline |
Synchronize with concurrent execution on the default stream.
When running with a GPU programming model, wait for currently active work on the primary stream or queue is finished.