AssemblySchedule Struct Reference#
|
DiFfRG
Discretization Framework for functional Renormalization Group flows
|
DiFfRG::AssemblySchedule Struct Reference
The two trailing arguments of dealii::MeshWorker::mesh_loop. More...
#include <assembly_schedule.hh>
Public Attributes | |
| uint | queue_length = 1 |
| mesh_loop's queue_length: how many work items are live in the pipeline at once. | |
| uint | chunk_size = 1 |
| mesh_loop's chunk_size: cells per work item, processed back-to-back by one worker. | |
Friends | |
| bool | operator== (const AssemblySchedule &, const AssemblySchedule &)=default |
Detailed Description
The two trailing arguments of dealii::MeshWorker::mesh_loop.
Neither value can change a computed number: mesh_loop drives a tbb::parallel_pipeline whose copier stage is serial and runs in iterator order, so the assembled result is bitwise identical for any schedule. They are purely a performance choice, which is what makes it safe to derive them rather than pin them per configuration.
Friends And Related Symbol Documentation
◆ operator==
|
friend |
Member Data Documentation
◆ chunk_size
| uint DiFfRG::AssemblySchedule::chunk_size = 1 |
mesh_loop's chunk_size: cells per work item, processed back-to-back by one worker.
◆ queue_length
| uint DiFfRG::AssemblySchedule::queue_length = 1 |
mesh_loop's queue_length: how many work items are live in the pipeline at once.
The documentation for this struct was generated from the following file:
- /home/runner/work/DiFfRG_current/DiFfRG_current/DiFfRG/include/DiFfRG/discretization/common/assembly_schedule.hh
Generated by