Visible to Intel only — GUID: GUID-979E1164-4514-4812-9A73-2F81510F00F5
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
Invoke a Callable Object
Appendix A Costs of Time Slicing
Appendix B Mixing With Other Threading Packages
References
parallel_for_each Body semantics and requirements
parallel_sort ranges interface extension
TBB_malloc_replacement_log Function
Type-specified message keys for join_node
Scalable Memory Pools
Helper Functions for Expressing Graphs
concurrent_lru_cache
task_group extensions
The customizing mutex type for concurrent_hash_map
Visible to Intel only — GUID: GUID-979E1164-4514-4812-9A73-2F81510F00F5
Flow Graph Basics: Edges
Most applications contain multiple nodes with edges connecting them to each other. In the flow graph interface, edges are directed channels over which messages are passed. They are created by calling the function make_edge( p, s ) with two arguments: p, the predecessor, and s, the successor. You can modify the example used in the Nodes topic to include a second node that squares the value it receives before printing it and then connect that to the first node with an edge.
graph g;
function_node< int, int > n( g, unlimited, []( int v ) -> int {
cout << v;
spin_for( v );
cout << v;
return v;
} );
function_node< int, int > m( g, 1, []( int v ) -> int {
v *= v;
cout << v;
spin_for( v );
cout << v;
return v;
} );
make_edge( n, m );
n.try_put( 1 );
n.try_put( 2 );
n.try_put( 3 );
g.wait_for_all();Now there are two function_nodes, n and m. The call to make_edge creates an edge from n to m. The node n is created with unlimited concurrency, while m has a concurrency limit of 1. The invocations of n can all proceed in parallel, while the invocations of m will be serialized. Because there is an edge from n to m, each value v, returned by n, will be automatically passed to node m by the runtime library.