Trait rayon::par_iter::ParallelIterator [−] [src]

pub trait ParallelIterator: Sized {
    type Item: Send;
    fn weight(self, scale: f64) -> Weight<Self> { ... }
    fn weight_max(self) -> Weight<Self> { ... }
    fn for_each<OP>(self, op: OP) where OP: Fn(Self::Item) + Sync { ... }
    fn map<MAP_OP, R>(self, map_op: MAP_OP) -> Map<Self, MapFn<MAP_OP>> where MAP_OP: Fn(Self::Item) -> R + Sync { ... }
    fn cloned<'a, T>(self) -> Map<Self, MapCloned> where T: 'a + Clone, Self: ParallelIterator<Item=&'a T> { ... }
    fn inspect<INSPECT_OP>(self, inspect_op: INSPECT_OP) -> Map<Self, MapInspect<INSPECT_OP>> where INSPECT_OP: Fn(&Self::Item) + Sync { ... }
    fn filter<FILTER_OP>(self, filter_op: FILTER_OP) -> Filter<Self, FILTER_OP> where FILTER_OP: Fn(&Self::Item) -> bool + Sync { ... }
    fn filter_map<FILTER_OP, R>(self, filter_op: FILTER_OP) -> FilterMap<Self, FILTER_OP> where FILTER_OP: Fn(Self::Item) -> Option<R> + Sync { ... }
    fn flat_map<MAP_OP, PI>(self, map_op: MAP_OP) -> FlatMap<Self, MAP_OP> where MAP_OP: Fn(Self::Item) -> PI + Sync, PI: IntoParallelIterator { ... }
    fn reduce_with<OP>(self, op: OP) -> Option<Self::Item> where OP: Fn(Self::Item, Self::Item) -> Self::Item + Sync { ... }
    fn reduce_with_identity<OP>(self, identity: Self::Item, op: OP) -> Self::Item where OP: Fn(Self::Item, Self::Item) -> Self::Item + Sync, Self::Item: Clone + Sync { ... }
    fn sum(self) -> Self::Item where SumOp: ReduceOp<Self::Item> { ... }
    fn mul(self) -> Self::Item where MulOp: ReduceOp<Self::Item> { ... }
    fn min(self) -> Self::Item where MinOp: ReduceOp<Self::Item> { ... }
    fn max(self) -> Self::Item where MaxOp: ReduceOp<Self::Item> { ... }
    fn reduce<REDUCE_OP>(self, reduce_op: &REDUCE_OP) -> Self::Item where REDUCE_OP: ReduceOp<Self::Item> { ... }
}

The ParallelIterator interface.

Associated Types

type Item: Send

Provided Methods

fn weight(self, scale: f64) -> Weight<Self>

Indicates the relative "weight" of producing each item in this parallel iterator. A higher weight will cause finer-grained parallel subtasks. 1.0 indicates something very cheap and uniform, like copying a value out of an array, or computing x + 1. If your tasks are either very expensive, or very unpredictable, you are better off with higher values. See also weight_max, which is a convenient shorthand to force the finest grained parallel execution posible. Tuning this value should not affect correctness but can improve (or hurt) performance.

fn weight_max(self) -> Weight<Self>

Shorthand for self.weight(f64::INFINITY). This forces the smallest granularity of parallel execution, which makes sense when your parallel tasks are (potentially) very expensive to execute.

fn for_each<OP>(self, op: OP) where OP: Fn(Self::Item) + Sync

Executes OP on each item produced by the iterator, in parallel.

fn map<MAP_OP, R>(self, map_op: MAP_OP) -> Map<Self, MapFn<MAP_OP>> where MAP_OP: Fn(Self::Item) -> R + Sync

Applies map_op to each item of this iterator, producing a new iterator with the results.

fn cloned<'a, T>(self) -> Map<Self, MapCloned> where T: 'a + Clone, Self: ParallelIterator<Item=&'a T>

Creates an iterator which clones all of its elements. This may be useful when you have an iterator over &T, but you need T.

fn inspect<INSPECT_OP>(self, inspect_op: INSPECT_OP) -> Map<Self, MapInspect<INSPECT_OP>> where INSPECT_OP: Fn(&Self::Item) + Sync

Applies inspect_op to a reference to each item of this iterator, producing a new iterator passing through the original items. This is often useful for debugging to see what's happening in iterator stages.

fn filter<FILTER_OP>(self, filter_op: FILTER_OP) -> Filter<Self, FILTER_OP> where FILTER_OP: Fn(&Self::Item) -> bool + Sync

Applies filter_op to each item of this iterator, producing a new iterator with only the items that gave true results.

fn filter_map<FILTER_OP, R>(self, filter_op: FILTER_OP) -> FilterMap<Self, FILTER_OP> where FILTER_OP: Fn(Self::Item) -> Option<R> + Sync

Applies filter_op to each item of this iterator to get an Option, producing a new iterator with only the items from Some results.

fn flat_map<MAP_OP, PI>(self, map_op: MAP_OP) -> FlatMap<Self, MAP_OP> where MAP_OP: Fn(Self::Item) -> PI + Sync, PI: IntoParallelIterator

Applies map_op to each item of this iterator to get nested iterators, producing a new iterator that flattens these back into one.

fn reduce_with<OP>(self, op: OP) -> Option<Self::Item> where OP: Fn(Self::Item, Self::Item) -> Self::Item + Sync

Reduces the items in the iterator into one item using op. See also sum, mul, min, etc, which are slightly more efficient. Returns None if the iterator is empty.

Note: unlike in a sequential iterator, the order in which op will be applied to reduce the result is not specified. So op should be commutative and associative or else the results will be non-deterministic.

fn reduce_with_identity<OP>(self, identity: Self::Item, op: OP) -> Self::Item where OP: Fn(Self::Item, Self::Item) -> Self::Item + Sync, Self::Item: Clone + Sync

Reduces the items in the iterator into one item using op. The argument identity represents an "identity" value which may be inserted into the sequence as needed to create opportunities for parallel execution. So, for example, if you are doing a summation, then identity ought to be something that represents the zero for your type (but consider just calling sum() in that case).

Example vectors.par_iter().reduce_with_identity(Vector::zero(), Vector::add).

Note: unlike in a sequential iterator, the order in which op will be applied to reduce the result is not specified. So op should be commutative and associative or else the results will be non-deterministic. And of course identity should be a true identity.

fn sum(self) -> Self::Item where SumOp: ReduceOp<Self::Item>

Sums up the items in the iterator.

Note that the order in items will be reduced is not specified, so if the + operator is not truly commutative and associative (as is the case for floating point numbers), then the results are not fully deterministic.

fn mul(self) -> Self::Item where MulOp: ReduceOp<Self::Item>

Multiplies all the items in the iterator.

Note that the order in items will be reduced is not specified, so if the * operator is not truly commutative and associative (as is the case for floating point numbers), then the results are not fully deterministic.

fn min(self) -> Self::Item where MinOp: ReduceOp<Self::Item>

Computes the minimum of all the items in the iterator.

Note that the order in items will be reduced is not specified, so if the Ord impl is not truly commutative and associative (as is the case for floating point numbers), then the results are not deterministic.

fn max(self) -> Self::Item where MaxOp: ReduceOp<Self::Item>

Computes the maximum of all the items in the iterator.

Note that the order in items will be reduced is not specified, so if the Ord impl is not truly commutative and associative (as is the case for floating point numbers), then the results are not deterministic.

fn reduce<REDUCE_OP>(self, reduce_op: &REDUCE_OP) -> Self::Item where REDUCE_OP: ReduceOp<Self::Item>

Reduces the items using the given "reduce operator". You may prefer reduce_with for a simpler interface.

Note that the order in items will be reduced is not specified, so if the reduce_op impl is not truly commutative and associative, then the results are not deterministic.

Implementors