Visible to Intel only — GUID: GUID-63875387-A85F-415B-A91F-491EAB28277F
Package Contents
Parallelizing Simple Loops
Parallelizing Complex Loops
Parallelizing Data Flow and Dependence Graphs
Work Isolation
Exceptions and Cancellation
Containers
Mutual Exclusion
Timing
Memory Allocation
The Task Scheduler
Design Patterns
Migrating from Threading Building Blocks (TBB)
Constrained APIs
Appendix A Costs of Time Slicing
Appendix B Mixing With Other Threading Packages
References
oneapi::tbb::info namespace
parallel_for_each Body semantics and requirements
parallel_sort ranges interface extension
Type-specified message keys for join_node
Scalable Memory Pools
Helper Functions for Expressing Graphs
concurrent_lru_cache
task_arena::constraints extensions
oneapi::tbb::info namespace extensions
task_group extensions
The customizing mutex type for concurrent_hash_map
Visible to Intel only — GUID: GUID-63875387-A85F-415B-A91F-491EAB28277F
Attach Flow Graph to an Arbitrary Task Arena
oneTBB task_arena interface provides mechanisms to guide tasks execution within the arena by setting the preferred computation units, restricting part of computation units, or limiting arena concurrency. In some cases, you may want to apply such mechanisms when a flow graph executes.
During its construction, a graph object attaches to the arena, in which the constructing thread occupies a slot.
This example shows how to set the most performant core type as the preferred one for a graph execution:
std::vector<tbb::core_type_id> core_types = tbb::info::core_types();
tbb::task_arena arena(
tbb::task_arena::constraints{}.set_core_type(core_types.back())
);
arena.execute( [&]() {
graph g;
function_node< int > f( g, unlimited, []( int ) {
/*the most performant core type is defined as preferred.*/
} );
f.try_put(1);
g.wait_for_all();
} );
A graph object can be reattached to a different task_arena by calling the graph::reset() function. It reinitializes and reattaches the graph to the task arena instance, inside which the graph::reset() method is executed.
This example shows how to reattach existing graph to an arena with the most performant core type as the preferred one for a work execution. Whenever a task is spawned on behalf of the graph, it is spawned in the arena of a graph it is attached to, disregarding the arena of the thread that the task is spawned from:
graph g;
function_node< int > f( g, unlimited, []( int ) {
/*the most performant core type is defined as preferred.*/
} );
std::vector<tbb::core_type_id> core_types = tbb::info::core_types();
tbb::task_arena arena(
tbb::task_arena::constraints{}.set_core_type(core_types.back())
);
arena.execute( [&]() {
g.reset();
} );
f.try_put(1);
g.wait_for_all();
See the following topics to learn more: