tmp/tmpv2v70oqv/{from.md → to.md}
RENAMED
|
@@ -3,22 +3,31 @@
|
|
| 3 |
Subclause [[algorithms.parallel]] describes components that C++ programs
|
| 4 |
may use to perform operations on containers and other sequences in
|
| 5 |
parallel.
|
| 6 |
|
| 7 |
A *parallel algorithm* is a function template listed in this document
|
| 8 |
-
with a template parameter named `ExecutionPolicy`
|
|
|
|
| 9 |
|
| 10 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 11 |
arguments by invoking the following functions:
|
| 12 |
|
| 13 |
-
- All operations of the categories of the iterators
|
| 14 |
-
is instantiated with.
|
| 15 |
- Operations on those sequence elements that are required by its
|
| 16 |
specification.
|
| 17 |
-
- User-provided
|
| 18 |
the algorithm, if required by the specification.
|
| 19 |
-
- Operations on those
|
| 20 |
\[*Note 1*: See [[algorithms.requirements]]. — *end note*]
|
| 21 |
|
| 22 |
These functions are herein called *element access functions*.
|
| 23 |
|
| 24 |
[*Example 1*:
|
|
@@ -35,11 +44,12 @@ The `sort` function may invoke the following element access functions:
|
|
| 35 |
— *end example*]
|
| 36 |
|
| 37 |
A standard library function is *vectorization-unsafe* if it is specified
|
| 38 |
to synchronize with another function invocation, or another function
|
| 39 |
invocation is specified to synchronize with it, and if it is not a
|
| 40 |
-
memory allocation or deallocation function
|
|
|
|
| 41 |
|
| 42 |
[*Note 2*: Implementations must ensure that internal synchronization
|
| 43 |
inside standard library functions does not prevent forward progress when
|
| 44 |
those functions are executed by threads of execution with weakly
|
| 45 |
parallel forward progress guarantees. — *end note*]
|
|
|
|
| 3 |
Subclause [[algorithms.parallel]] describes components that C++ programs
|
| 4 |
may use to perform operations on containers and other sequences in
|
| 5 |
parallel.
|
| 6 |
|
| 7 |
A *parallel algorithm* is a function template listed in this document
|
| 8 |
+
with a template parameter named `ExecutionPolicy` or constrained by the
|
| 9 |
+
following exposition-only concept:
|
| 10 |
|
| 11 |
+
``` cpp
|
| 12 |
+
template<class Ep>
|
| 13 |
+
concept execution-policy = is_execution_policy_v<remove_cvref_t<Ep>>; // exposition only
|
| 14 |
+
```
|
| 15 |
+
|
| 16 |
+
Such a template parameter is termed an
|
| 17 |
+
*execution policy template parameter*.
|
| 18 |
+
|
| 19 |
+
A parallel algorithm accesses objects indirectly accessible via its
|
| 20 |
arguments by invoking the following functions:
|
| 21 |
|
| 22 |
+
- All operations of the categories of the iterators, sentinels, or
|
| 23 |
+
`mdspan` types that the algorithm is instantiated with.
|
| 24 |
- Operations on those sequence elements that are required by its
|
| 25 |
specification.
|
| 26 |
+
- User-provided invocable objects to be applied during the execution of
|
| 27 |
the algorithm, if required by the specification.
|
| 28 |
+
- Operations on those invocable objects required by the specification.
|
| 29 |
\[*Note 1*: See [[algorithms.requirements]]. — *end note*]
|
| 30 |
|
| 31 |
These functions are herein called *element access functions*.
|
| 32 |
|
| 33 |
[*Example 1*:
|
|
|
|
| 44 |
— *end example*]
|
| 45 |
|
| 46 |
A standard library function is *vectorization-unsafe* if it is specified
|
| 47 |
to synchronize with another function invocation, or another function
|
| 48 |
invocation is specified to synchronize with it, and if it is not a
|
| 49 |
+
memory allocation or deallocation function or lock-free atomic
|
| 50 |
+
modify-write operation [[atomics.order]].
|
| 51 |
|
| 52 |
[*Note 2*: Implementations must ensure that internal synchronization
|
| 53 |
inside standard library functions does not prevent forward progress when
|
| 54 |
those functions are executed by threads of execution with weakly
|
| 55 |
parallel forward progress guarantees. — *end note*]
|