2017-02-03 17:04:22 -06:00
|
|
|
use core::iter::*;
|
|
|
|
use test::{Bencher, black_box};
|
|
|
|
|
|
|
|
#[bench]
|
|
|
|
fn bench_rposition(b: &mut Bencher) {
|
|
|
|
let it: Vec<usize> = (0..300).collect();
|
|
|
|
b.iter(|| {
|
|
|
|
it.iter().rposition(|&x| x <= 150);
|
|
|
|
});
|
|
|
|
}
|
|
|
|
|
|
|
|
#[bench]
|
|
|
|
fn bench_skip_while(b: &mut Bencher) {
|
|
|
|
b.iter(|| {
|
|
|
|
let it = 0..100;
|
|
|
|
let mut sum = 0;
|
|
|
|
it.skip_while(|&x| { sum += x; sum < 4000 }).all(|_| true);
|
|
|
|
});
|
|
|
|
}
|
|
|
|
|
|
|
|
#[bench]
|
|
|
|
fn bench_multiple_take(b: &mut Bencher) {
|
|
|
|
let mut it = (0..42).cycle();
|
|
|
|
b.iter(|| {
|
|
|
|
let n = it.next().unwrap();
|
|
|
|
for _ in 0..n {
|
|
|
|
it.clone().take(it.next().unwrap()).all(|_| true);
|
|
|
|
}
|
|
|
|
});
|
|
|
|
}
|
|
|
|
|
|
|
|
fn scatter(x: i32) -> i32 { (x * 31) % 127 }
|
|
|
|
|
|
|
|
#[bench]
|
|
|
|
fn bench_max_by_key(b: &mut Bencher) {
|
|
|
|
b.iter(|| {
|
|
|
|
let it = 0..100;
|
|
|
|
it.max_by_key(|&x| scatter(x))
|
|
|
|
})
|
|
|
|
}
|
|
|
|
|
|
|
|
// http://www.reddit.com/r/rust/comments/31syce/using_iterators_to_find_the_index_of_the_min_or/
|
|
|
|
#[bench]
|
|
|
|
fn bench_max_by_key2(b: &mut Bencher) {
|
|
|
|
fn max_index_iter(array: &[i32]) -> usize {
|
|
|
|
array.iter().enumerate().max_by_key(|&(_, item)| item).unwrap().0
|
|
|
|
}
|
|
|
|
|
|
|
|
let mut data = vec![0; 1638];
|
|
|
|
data[514] = 9999;
|
|
|
|
|
|
|
|
b.iter(|| max_index_iter(&data));
|
|
|
|
}
|
|
|
|
|
|
|
|
#[bench]
|
|
|
|
fn bench_max(b: &mut Bencher) {
|
|
|
|
b.iter(|| {
|
|
|
|
let it = 0..100;
|
|
|
|
it.map(scatter).max()
|
|
|
|
})
|
|
|
|
}
|
|
|
|
|
|
|
|
pub fn copy_zip(xs: &[u8], ys: &mut [u8]) {
|
|
|
|
for (a, b) in ys.iter_mut().zip(xs) {
|
|
|
|
*a = *b;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
pub fn add_zip(xs: &[f32], ys: &mut [f32]) {
|
|
|
|
for (a, b) in ys.iter_mut().zip(xs) {
|
|
|
|
*a += *b;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
#[bench]
|
|
|
|
fn bench_zip_copy(b: &mut Bencher) {
|
|
|
|
let source = vec![0u8; 16 * 1024];
|
|
|
|
let mut dst = black_box(vec![0u8; 16 * 1024]);
|
|
|
|
b.iter(|| {
|
|
|
|
copy_zip(&source, &mut dst)
|
|
|
|
})
|
|
|
|
}
|
|
|
|
|
|
|
|
#[bench]
|
|
|
|
fn bench_zip_add(b: &mut Bencher) {
|
|
|
|
let source = vec![1.; 16 * 1024];
|
|
|
|
let mut dst = vec![0.; 16 * 1024];
|
|
|
|
b.iter(|| {
|
|
|
|
add_zip(&source, &mut dst)
|
|
|
|
});
|
|
|
|
}
|
2017-06-21 15:22:27 -05:00
|
|
|
|
|
|
|
/// `Iterator::for_each` implemented as a plain loop.
|
|
|
|
fn for_each_loop<I, F>(iter: I, mut f: F) where
|
|
|
|
I: Iterator, F: FnMut(I::Item)
|
|
|
|
{
|
|
|
|
for item in iter {
|
|
|
|
f(item);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/// `Iterator::for_each` implemented with `fold` for internal iteration.
|
|
|
|
/// (except when `by_ref()` effectively disables that optimization.)
|
|
|
|
fn for_each_fold<I, F>(iter: I, mut f: F) where
|
|
|
|
I: Iterator, F: FnMut(I::Item)
|
|
|
|
{
|
|
|
|
iter.fold((), move |(), item| f(item));
|
|
|
|
}
|
|
|
|
|
|
|
|
#[bench]
|
|
|
|
fn bench_for_each_chain_loop(b: &mut Bencher) {
|
|
|
|
b.iter(|| {
|
|
|
|
let mut acc = 0;
|
|
|
|
let iter = (0i64..1000000).chain(0..1000000).map(black_box);
|
|
|
|
for_each_loop(iter, |x| acc += x);
|
|
|
|
acc
|
|
|
|
});
|
|
|
|
}
|
|
|
|
|
|
|
|
#[bench]
|
|
|
|
fn bench_for_each_chain_fold(b: &mut Bencher) {
|
|
|
|
b.iter(|| {
|
|
|
|
let mut acc = 0;
|
|
|
|
let iter = (0i64..1000000).chain(0..1000000).map(black_box);
|
|
|
|
for_each_fold(iter, |x| acc += x);
|
|
|
|
acc
|
|
|
|
});
|
|
|
|
}
|
|
|
|
|
|
|
|
#[bench]
|
|
|
|
fn bench_for_each_chain_ref_fold(b: &mut Bencher) {
|
|
|
|
b.iter(|| {
|
|
|
|
let mut acc = 0;
|
|
|
|
let mut iter = (0i64..1000000).chain(0..1000000).map(black_box);
|
|
|
|
for_each_fold(iter.by_ref(), |x| acc += x);
|
|
|
|
acc
|
|
|
|
});
|
|
|
|
}
|
Customize `<FlatMap as Iterator>::fold`
`FlatMap` can use internal iteration for its `fold`, which shows a
performance advantage in the new benchmarks:
test iter::bench_flat_map_chain_ref_sum ... bench: 4,354,111 ns/iter (+/- 108,871)
test iter::bench_flat_map_chain_sum ... bench: 468,167 ns/iter (+/- 2,274)
test iter::bench_flat_map_ref_sum ... bench: 449,616 ns/iter (+/- 6,257)
test iter::bench_flat_map_sum ... bench: 348,010 ns/iter (+/- 1,227)
... where the "ref" benches are using `by_ref()` that isn't optimized.
So this change shows a decent advantage on its own, but much more when
combined with a `chain` iterator that also optimizes `fold`.
2017-09-14 15:51:32 -05:00
|
|
|
|
Add more custom folding to `core::iter` adaptors
Many of the iterator adaptors will perform faster folds if they forward
to their inner iterator's folds, especially for inner types like `Chain`
which are optimized too. The following types are newly specialized:
| Type | `fold` | `rfold` |
| ----------- | ------ | ------- |
| `Enumerate` | ✓ | ✓ |
| `Filter` | ✓ | ✓ |
| `FilterMap` | ✓ | ✓ |
| `FlatMap` | exists | ✓ |
| `Fuse` | ✓ | ✓ |
| `Inspect` | ✓ | ✓ |
| `Peekable` | ✓ | N/A¹ |
| `Skip` | ✓ | N/A² |
| `SkipWhile` | ✓ | N/A¹ |
¹ not a `DoubleEndedIterator`
² `Skip::next_back` doesn't pull skipped items at all, but this couldn't
be avoided if `Skip::rfold` were to call its inner iterator's `rfold`.
Benchmarks
----------
In the following results, plain `_sum` computes the sum of a million
integers -- note that `sum()` is implemented with `fold()`. The
`_ref_sum` variants do the same on a `by_ref()` iterator, which is
limited to calling `next()` one by one, without specialized `fold`.
The `chain` variants perform the same tests on two iterators chained
together, to show a greater benefit of forwarding `fold` internally.
test iter::bench_enumerate_chain_ref_sum ... bench: 2,216,264 ns/iter (+/- 29,228)
test iter::bench_enumerate_chain_sum ... bench: 922,380 ns/iter (+/- 2,676)
test iter::bench_enumerate_ref_sum ... bench: 476,094 ns/iter (+/- 7,110)
test iter::bench_enumerate_sum ... bench: 476,438 ns/iter (+/- 3,334)
test iter::bench_filter_chain_ref_sum ... bench: 2,266,095 ns/iter (+/- 6,051)
test iter::bench_filter_chain_sum ... bench: 745,594 ns/iter (+/- 2,013)
test iter::bench_filter_ref_sum ... bench: 889,696 ns/iter (+/- 1,188)
test iter::bench_filter_sum ... bench: 667,325 ns/iter (+/- 1,894)
test iter::bench_filter_map_chain_ref_sum ... bench: 2,259,195 ns/iter (+/- 353,440)
test iter::bench_filter_map_chain_sum ... bench: 1,223,280 ns/iter (+/- 1,972)
test iter::bench_filter_map_ref_sum ... bench: 611,607 ns/iter (+/- 2,507)
test iter::bench_filter_map_sum ... bench: 611,610 ns/iter (+/- 472)
test iter::bench_fuse_chain_ref_sum ... bench: 2,246,106 ns/iter (+/- 22,395)
test iter::bench_fuse_chain_sum ... bench: 634,887 ns/iter (+/- 1,341)
test iter::bench_fuse_ref_sum ... bench: 444,816 ns/iter (+/- 1,748)
test iter::bench_fuse_sum ... bench: 316,954 ns/iter (+/- 2,616)
test iter::bench_inspect_chain_ref_sum ... bench: 2,245,431 ns/iter (+/- 21,371)
test iter::bench_inspect_chain_sum ... bench: 631,645 ns/iter (+/- 4,928)
test iter::bench_inspect_ref_sum ... bench: 317,437 ns/iter (+/- 702)
test iter::bench_inspect_sum ... bench: 315,942 ns/iter (+/- 4,320)
test iter::bench_peekable_chain_ref_sum ... bench: 2,243,585 ns/iter (+/- 12,186)
test iter::bench_peekable_chain_sum ... bench: 634,848 ns/iter (+/- 1,712)
test iter::bench_peekable_ref_sum ... bench: 444,808 ns/iter (+/- 480)
test iter::bench_peekable_sum ... bench: 317,133 ns/iter (+/- 3,309)
test iter::bench_skip_chain_ref_sum ... bench: 1,778,734 ns/iter (+/- 2,198)
test iter::bench_skip_chain_sum ... bench: 761,850 ns/iter (+/- 1,645)
test iter::bench_skip_ref_sum ... bench: 478,207 ns/iter (+/- 119,252)
test iter::bench_skip_sum ... bench: 315,614 ns/iter (+/- 3,054)
test iter::bench_skip_while_chain_ref_sum ... bench: 2,486,370 ns/iter (+/- 4,845)
test iter::bench_skip_while_chain_sum ... bench: 633,915 ns/iter (+/- 5,892)
test iter::bench_skip_while_ref_sum ... bench: 666,926 ns/iter (+/- 804)
test iter::bench_skip_while_sum ... bench: 444,405 ns/iter (+/- 571)
2017-09-25 22:53:08 -05:00
|
|
|
|
|
|
|
/// Helper to benchmark `sum` for iterators taken by value which
|
|
|
|
/// can optimize `fold`, and by reference which cannot.
|
|
|
|
macro_rules! bench_sums {
|
|
|
|
($bench_sum:ident, $bench_ref_sum:ident, $iter:expr) => {
|
|
|
|
#[bench]
|
|
|
|
fn $bench_sum(b: &mut Bencher) {
|
|
|
|
b.iter(|| -> i64 {
|
|
|
|
$iter.map(black_box).sum()
|
|
|
|
});
|
|
|
|
}
|
|
|
|
|
|
|
|
#[bench]
|
|
|
|
fn $bench_ref_sum(b: &mut Bencher) {
|
|
|
|
b.iter(|| -> i64 {
|
|
|
|
$iter.map(black_box).by_ref().sum()
|
|
|
|
});
|
|
|
|
}
|
|
|
|
}
|
Customize `<FlatMap as Iterator>::fold`
`FlatMap` can use internal iteration for its `fold`, which shows a
performance advantage in the new benchmarks:
test iter::bench_flat_map_chain_ref_sum ... bench: 4,354,111 ns/iter (+/- 108,871)
test iter::bench_flat_map_chain_sum ... bench: 468,167 ns/iter (+/- 2,274)
test iter::bench_flat_map_ref_sum ... bench: 449,616 ns/iter (+/- 6,257)
test iter::bench_flat_map_sum ... bench: 348,010 ns/iter (+/- 1,227)
... where the "ref" benches are using `by_ref()` that isn't optimized.
So this change shows a decent advantage on its own, but much more when
combined with a `chain` iterator that also optimizes `fold`.
2017-09-14 15:51:32 -05:00
|
|
|
}
|
|
|
|
|
Add more custom folding to `core::iter` adaptors
Many of the iterator adaptors will perform faster folds if they forward
to their inner iterator's folds, especially for inner types like `Chain`
which are optimized too. The following types are newly specialized:
| Type | `fold` | `rfold` |
| ----------- | ------ | ------- |
| `Enumerate` | ✓ | ✓ |
| `Filter` | ✓ | ✓ |
| `FilterMap` | ✓ | ✓ |
| `FlatMap` | exists | ✓ |
| `Fuse` | ✓ | ✓ |
| `Inspect` | ✓ | ✓ |
| `Peekable` | ✓ | N/A¹ |
| `Skip` | ✓ | N/A² |
| `SkipWhile` | ✓ | N/A¹ |
¹ not a `DoubleEndedIterator`
² `Skip::next_back` doesn't pull skipped items at all, but this couldn't
be avoided if `Skip::rfold` were to call its inner iterator's `rfold`.
Benchmarks
----------
In the following results, plain `_sum` computes the sum of a million
integers -- note that `sum()` is implemented with `fold()`. The
`_ref_sum` variants do the same on a `by_ref()` iterator, which is
limited to calling `next()` one by one, without specialized `fold`.
The `chain` variants perform the same tests on two iterators chained
together, to show a greater benefit of forwarding `fold` internally.
test iter::bench_enumerate_chain_ref_sum ... bench: 2,216,264 ns/iter (+/- 29,228)
test iter::bench_enumerate_chain_sum ... bench: 922,380 ns/iter (+/- 2,676)
test iter::bench_enumerate_ref_sum ... bench: 476,094 ns/iter (+/- 7,110)
test iter::bench_enumerate_sum ... bench: 476,438 ns/iter (+/- 3,334)
test iter::bench_filter_chain_ref_sum ... bench: 2,266,095 ns/iter (+/- 6,051)
test iter::bench_filter_chain_sum ... bench: 745,594 ns/iter (+/- 2,013)
test iter::bench_filter_ref_sum ... bench: 889,696 ns/iter (+/- 1,188)
test iter::bench_filter_sum ... bench: 667,325 ns/iter (+/- 1,894)
test iter::bench_filter_map_chain_ref_sum ... bench: 2,259,195 ns/iter (+/- 353,440)
test iter::bench_filter_map_chain_sum ... bench: 1,223,280 ns/iter (+/- 1,972)
test iter::bench_filter_map_ref_sum ... bench: 611,607 ns/iter (+/- 2,507)
test iter::bench_filter_map_sum ... bench: 611,610 ns/iter (+/- 472)
test iter::bench_fuse_chain_ref_sum ... bench: 2,246,106 ns/iter (+/- 22,395)
test iter::bench_fuse_chain_sum ... bench: 634,887 ns/iter (+/- 1,341)
test iter::bench_fuse_ref_sum ... bench: 444,816 ns/iter (+/- 1,748)
test iter::bench_fuse_sum ... bench: 316,954 ns/iter (+/- 2,616)
test iter::bench_inspect_chain_ref_sum ... bench: 2,245,431 ns/iter (+/- 21,371)
test iter::bench_inspect_chain_sum ... bench: 631,645 ns/iter (+/- 4,928)
test iter::bench_inspect_ref_sum ... bench: 317,437 ns/iter (+/- 702)
test iter::bench_inspect_sum ... bench: 315,942 ns/iter (+/- 4,320)
test iter::bench_peekable_chain_ref_sum ... bench: 2,243,585 ns/iter (+/- 12,186)
test iter::bench_peekable_chain_sum ... bench: 634,848 ns/iter (+/- 1,712)
test iter::bench_peekable_ref_sum ... bench: 444,808 ns/iter (+/- 480)
test iter::bench_peekable_sum ... bench: 317,133 ns/iter (+/- 3,309)
test iter::bench_skip_chain_ref_sum ... bench: 1,778,734 ns/iter (+/- 2,198)
test iter::bench_skip_chain_sum ... bench: 761,850 ns/iter (+/- 1,645)
test iter::bench_skip_ref_sum ... bench: 478,207 ns/iter (+/- 119,252)
test iter::bench_skip_sum ... bench: 315,614 ns/iter (+/- 3,054)
test iter::bench_skip_while_chain_ref_sum ... bench: 2,486,370 ns/iter (+/- 4,845)
test iter::bench_skip_while_chain_sum ... bench: 633,915 ns/iter (+/- 5,892)
test iter::bench_skip_while_ref_sum ... bench: 666,926 ns/iter (+/- 804)
test iter::bench_skip_while_sum ... bench: 444,405 ns/iter (+/- 571)
2017-09-25 22:53:08 -05:00
|
|
|
bench_sums! {
|
|
|
|
bench_flat_map_sum,
|
|
|
|
bench_flat_map_ref_sum,
|
|
|
|
(0i64..1000).flat_map(|x| x..x+1000)
|
Customize `<FlatMap as Iterator>::fold`
`FlatMap` can use internal iteration for its `fold`, which shows a
performance advantage in the new benchmarks:
test iter::bench_flat_map_chain_ref_sum ... bench: 4,354,111 ns/iter (+/- 108,871)
test iter::bench_flat_map_chain_sum ... bench: 468,167 ns/iter (+/- 2,274)
test iter::bench_flat_map_ref_sum ... bench: 449,616 ns/iter (+/- 6,257)
test iter::bench_flat_map_sum ... bench: 348,010 ns/iter (+/- 1,227)
... where the "ref" benches are using `by_ref()` that isn't optimized.
So this change shows a decent advantage on its own, but much more when
combined with a `chain` iterator that also optimizes `fold`.
2017-09-14 15:51:32 -05:00
|
|
|
}
|
|
|
|
|
Add more custom folding to `core::iter` adaptors
Many of the iterator adaptors will perform faster folds if they forward
to their inner iterator's folds, especially for inner types like `Chain`
which are optimized too. The following types are newly specialized:
| Type | `fold` | `rfold` |
| ----------- | ------ | ------- |
| `Enumerate` | ✓ | ✓ |
| `Filter` | ✓ | ✓ |
| `FilterMap` | ✓ | ✓ |
| `FlatMap` | exists | ✓ |
| `Fuse` | ✓ | ✓ |
| `Inspect` | ✓ | ✓ |
| `Peekable` | ✓ | N/A¹ |
| `Skip` | ✓ | N/A² |
| `SkipWhile` | ✓ | N/A¹ |
¹ not a `DoubleEndedIterator`
² `Skip::next_back` doesn't pull skipped items at all, but this couldn't
be avoided if `Skip::rfold` were to call its inner iterator's `rfold`.
Benchmarks
----------
In the following results, plain `_sum` computes the sum of a million
integers -- note that `sum()` is implemented with `fold()`. The
`_ref_sum` variants do the same on a `by_ref()` iterator, which is
limited to calling `next()` one by one, without specialized `fold`.
The `chain` variants perform the same tests on two iterators chained
together, to show a greater benefit of forwarding `fold` internally.
test iter::bench_enumerate_chain_ref_sum ... bench: 2,216,264 ns/iter (+/- 29,228)
test iter::bench_enumerate_chain_sum ... bench: 922,380 ns/iter (+/- 2,676)
test iter::bench_enumerate_ref_sum ... bench: 476,094 ns/iter (+/- 7,110)
test iter::bench_enumerate_sum ... bench: 476,438 ns/iter (+/- 3,334)
test iter::bench_filter_chain_ref_sum ... bench: 2,266,095 ns/iter (+/- 6,051)
test iter::bench_filter_chain_sum ... bench: 745,594 ns/iter (+/- 2,013)
test iter::bench_filter_ref_sum ... bench: 889,696 ns/iter (+/- 1,188)
test iter::bench_filter_sum ... bench: 667,325 ns/iter (+/- 1,894)
test iter::bench_filter_map_chain_ref_sum ... bench: 2,259,195 ns/iter (+/- 353,440)
test iter::bench_filter_map_chain_sum ... bench: 1,223,280 ns/iter (+/- 1,972)
test iter::bench_filter_map_ref_sum ... bench: 611,607 ns/iter (+/- 2,507)
test iter::bench_filter_map_sum ... bench: 611,610 ns/iter (+/- 472)
test iter::bench_fuse_chain_ref_sum ... bench: 2,246,106 ns/iter (+/- 22,395)
test iter::bench_fuse_chain_sum ... bench: 634,887 ns/iter (+/- 1,341)
test iter::bench_fuse_ref_sum ... bench: 444,816 ns/iter (+/- 1,748)
test iter::bench_fuse_sum ... bench: 316,954 ns/iter (+/- 2,616)
test iter::bench_inspect_chain_ref_sum ... bench: 2,245,431 ns/iter (+/- 21,371)
test iter::bench_inspect_chain_sum ... bench: 631,645 ns/iter (+/- 4,928)
test iter::bench_inspect_ref_sum ... bench: 317,437 ns/iter (+/- 702)
test iter::bench_inspect_sum ... bench: 315,942 ns/iter (+/- 4,320)
test iter::bench_peekable_chain_ref_sum ... bench: 2,243,585 ns/iter (+/- 12,186)
test iter::bench_peekable_chain_sum ... bench: 634,848 ns/iter (+/- 1,712)
test iter::bench_peekable_ref_sum ... bench: 444,808 ns/iter (+/- 480)
test iter::bench_peekable_sum ... bench: 317,133 ns/iter (+/- 3,309)
test iter::bench_skip_chain_ref_sum ... bench: 1,778,734 ns/iter (+/- 2,198)
test iter::bench_skip_chain_sum ... bench: 761,850 ns/iter (+/- 1,645)
test iter::bench_skip_ref_sum ... bench: 478,207 ns/iter (+/- 119,252)
test iter::bench_skip_sum ... bench: 315,614 ns/iter (+/- 3,054)
test iter::bench_skip_while_chain_ref_sum ... bench: 2,486,370 ns/iter (+/- 4,845)
test iter::bench_skip_while_chain_sum ... bench: 633,915 ns/iter (+/- 5,892)
test iter::bench_skip_while_ref_sum ... bench: 666,926 ns/iter (+/- 804)
test iter::bench_skip_while_sum ... bench: 444,405 ns/iter (+/- 571)
2017-09-25 22:53:08 -05:00
|
|
|
bench_sums! {
|
|
|
|
bench_flat_map_chain_sum,
|
|
|
|
bench_flat_map_chain_ref_sum,
|
|
|
|
(0i64..1000000).flat_map(|x| once(x).chain(once(x)))
|
Customize `<FlatMap as Iterator>::fold`
`FlatMap` can use internal iteration for its `fold`, which shows a
performance advantage in the new benchmarks:
test iter::bench_flat_map_chain_ref_sum ... bench: 4,354,111 ns/iter (+/- 108,871)
test iter::bench_flat_map_chain_sum ... bench: 468,167 ns/iter (+/- 2,274)
test iter::bench_flat_map_ref_sum ... bench: 449,616 ns/iter (+/- 6,257)
test iter::bench_flat_map_sum ... bench: 348,010 ns/iter (+/- 1,227)
... where the "ref" benches are using `by_ref()` that isn't optimized.
So this change shows a decent advantage on its own, but much more when
combined with a `chain` iterator that also optimizes `fold`.
2017-09-14 15:51:32 -05:00
|
|
|
}
|
|
|
|
|
Add more custom folding to `core::iter` adaptors
Many of the iterator adaptors will perform faster folds if they forward
to their inner iterator's folds, especially for inner types like `Chain`
which are optimized too. The following types are newly specialized:
| Type | `fold` | `rfold` |
| ----------- | ------ | ------- |
| `Enumerate` | ✓ | ✓ |
| `Filter` | ✓ | ✓ |
| `FilterMap` | ✓ | ✓ |
| `FlatMap` | exists | ✓ |
| `Fuse` | ✓ | ✓ |
| `Inspect` | ✓ | ✓ |
| `Peekable` | ✓ | N/A¹ |
| `Skip` | ✓ | N/A² |
| `SkipWhile` | ✓ | N/A¹ |
¹ not a `DoubleEndedIterator`
² `Skip::next_back` doesn't pull skipped items at all, but this couldn't
be avoided if `Skip::rfold` were to call its inner iterator's `rfold`.
Benchmarks
----------
In the following results, plain `_sum` computes the sum of a million
integers -- note that `sum()` is implemented with `fold()`. The
`_ref_sum` variants do the same on a `by_ref()` iterator, which is
limited to calling `next()` one by one, without specialized `fold`.
The `chain` variants perform the same tests on two iterators chained
together, to show a greater benefit of forwarding `fold` internally.
test iter::bench_enumerate_chain_ref_sum ... bench: 2,216,264 ns/iter (+/- 29,228)
test iter::bench_enumerate_chain_sum ... bench: 922,380 ns/iter (+/- 2,676)
test iter::bench_enumerate_ref_sum ... bench: 476,094 ns/iter (+/- 7,110)
test iter::bench_enumerate_sum ... bench: 476,438 ns/iter (+/- 3,334)
test iter::bench_filter_chain_ref_sum ... bench: 2,266,095 ns/iter (+/- 6,051)
test iter::bench_filter_chain_sum ... bench: 745,594 ns/iter (+/- 2,013)
test iter::bench_filter_ref_sum ... bench: 889,696 ns/iter (+/- 1,188)
test iter::bench_filter_sum ... bench: 667,325 ns/iter (+/- 1,894)
test iter::bench_filter_map_chain_ref_sum ... bench: 2,259,195 ns/iter (+/- 353,440)
test iter::bench_filter_map_chain_sum ... bench: 1,223,280 ns/iter (+/- 1,972)
test iter::bench_filter_map_ref_sum ... bench: 611,607 ns/iter (+/- 2,507)
test iter::bench_filter_map_sum ... bench: 611,610 ns/iter (+/- 472)
test iter::bench_fuse_chain_ref_sum ... bench: 2,246,106 ns/iter (+/- 22,395)
test iter::bench_fuse_chain_sum ... bench: 634,887 ns/iter (+/- 1,341)
test iter::bench_fuse_ref_sum ... bench: 444,816 ns/iter (+/- 1,748)
test iter::bench_fuse_sum ... bench: 316,954 ns/iter (+/- 2,616)
test iter::bench_inspect_chain_ref_sum ... bench: 2,245,431 ns/iter (+/- 21,371)
test iter::bench_inspect_chain_sum ... bench: 631,645 ns/iter (+/- 4,928)
test iter::bench_inspect_ref_sum ... bench: 317,437 ns/iter (+/- 702)
test iter::bench_inspect_sum ... bench: 315,942 ns/iter (+/- 4,320)
test iter::bench_peekable_chain_ref_sum ... bench: 2,243,585 ns/iter (+/- 12,186)
test iter::bench_peekable_chain_sum ... bench: 634,848 ns/iter (+/- 1,712)
test iter::bench_peekable_ref_sum ... bench: 444,808 ns/iter (+/- 480)
test iter::bench_peekable_sum ... bench: 317,133 ns/iter (+/- 3,309)
test iter::bench_skip_chain_ref_sum ... bench: 1,778,734 ns/iter (+/- 2,198)
test iter::bench_skip_chain_sum ... bench: 761,850 ns/iter (+/- 1,645)
test iter::bench_skip_ref_sum ... bench: 478,207 ns/iter (+/- 119,252)
test iter::bench_skip_sum ... bench: 315,614 ns/iter (+/- 3,054)
test iter::bench_skip_while_chain_ref_sum ... bench: 2,486,370 ns/iter (+/- 4,845)
test iter::bench_skip_while_chain_sum ... bench: 633,915 ns/iter (+/- 5,892)
test iter::bench_skip_while_ref_sum ... bench: 666,926 ns/iter (+/- 804)
test iter::bench_skip_while_sum ... bench: 444,405 ns/iter (+/- 571)
2017-09-25 22:53:08 -05:00
|
|
|
bench_sums! {
|
|
|
|
bench_enumerate_sum,
|
|
|
|
bench_enumerate_ref_sum,
|
|
|
|
(0i64..1000000).enumerate().map(|(i, x)| x * i as i64)
|
|
|
|
}
|
|
|
|
|
|
|
|
bench_sums! {
|
|
|
|
bench_enumerate_chain_sum,
|
|
|
|
bench_enumerate_chain_ref_sum,
|
|
|
|
(0i64..1000000).chain(0..1000000).enumerate().map(|(i, x)| x * i as i64)
|
|
|
|
}
|
|
|
|
|
|
|
|
bench_sums! {
|
|
|
|
bench_filter_sum,
|
|
|
|
bench_filter_ref_sum,
|
2019-02-27 04:46:37 -06:00
|
|
|
(0i64..1000000).filter(|x| x % 3 == 0)
|
Add more custom folding to `core::iter` adaptors
Many of the iterator adaptors will perform faster folds if they forward
to their inner iterator's folds, especially for inner types like `Chain`
which are optimized too. The following types are newly specialized:
| Type | `fold` | `rfold` |
| ----------- | ------ | ------- |
| `Enumerate` | ✓ | ✓ |
| `Filter` | ✓ | ✓ |
| `FilterMap` | ✓ | ✓ |
| `FlatMap` | exists | ✓ |
| `Fuse` | ✓ | ✓ |
| `Inspect` | ✓ | ✓ |
| `Peekable` | ✓ | N/A¹ |
| `Skip` | ✓ | N/A² |
| `SkipWhile` | ✓ | N/A¹ |
¹ not a `DoubleEndedIterator`
² `Skip::next_back` doesn't pull skipped items at all, but this couldn't
be avoided if `Skip::rfold` were to call its inner iterator's `rfold`.
Benchmarks
----------
In the following results, plain `_sum` computes the sum of a million
integers -- note that `sum()` is implemented with `fold()`. The
`_ref_sum` variants do the same on a `by_ref()` iterator, which is
limited to calling `next()` one by one, without specialized `fold`.
The `chain` variants perform the same tests on two iterators chained
together, to show a greater benefit of forwarding `fold` internally.
test iter::bench_enumerate_chain_ref_sum ... bench: 2,216,264 ns/iter (+/- 29,228)
test iter::bench_enumerate_chain_sum ... bench: 922,380 ns/iter (+/- 2,676)
test iter::bench_enumerate_ref_sum ... bench: 476,094 ns/iter (+/- 7,110)
test iter::bench_enumerate_sum ... bench: 476,438 ns/iter (+/- 3,334)
test iter::bench_filter_chain_ref_sum ... bench: 2,266,095 ns/iter (+/- 6,051)
test iter::bench_filter_chain_sum ... bench: 745,594 ns/iter (+/- 2,013)
test iter::bench_filter_ref_sum ... bench: 889,696 ns/iter (+/- 1,188)
test iter::bench_filter_sum ... bench: 667,325 ns/iter (+/- 1,894)
test iter::bench_filter_map_chain_ref_sum ... bench: 2,259,195 ns/iter (+/- 353,440)
test iter::bench_filter_map_chain_sum ... bench: 1,223,280 ns/iter (+/- 1,972)
test iter::bench_filter_map_ref_sum ... bench: 611,607 ns/iter (+/- 2,507)
test iter::bench_filter_map_sum ... bench: 611,610 ns/iter (+/- 472)
test iter::bench_fuse_chain_ref_sum ... bench: 2,246,106 ns/iter (+/- 22,395)
test iter::bench_fuse_chain_sum ... bench: 634,887 ns/iter (+/- 1,341)
test iter::bench_fuse_ref_sum ... bench: 444,816 ns/iter (+/- 1,748)
test iter::bench_fuse_sum ... bench: 316,954 ns/iter (+/- 2,616)
test iter::bench_inspect_chain_ref_sum ... bench: 2,245,431 ns/iter (+/- 21,371)
test iter::bench_inspect_chain_sum ... bench: 631,645 ns/iter (+/- 4,928)
test iter::bench_inspect_ref_sum ... bench: 317,437 ns/iter (+/- 702)
test iter::bench_inspect_sum ... bench: 315,942 ns/iter (+/- 4,320)
test iter::bench_peekable_chain_ref_sum ... bench: 2,243,585 ns/iter (+/- 12,186)
test iter::bench_peekable_chain_sum ... bench: 634,848 ns/iter (+/- 1,712)
test iter::bench_peekable_ref_sum ... bench: 444,808 ns/iter (+/- 480)
test iter::bench_peekable_sum ... bench: 317,133 ns/iter (+/- 3,309)
test iter::bench_skip_chain_ref_sum ... bench: 1,778,734 ns/iter (+/- 2,198)
test iter::bench_skip_chain_sum ... bench: 761,850 ns/iter (+/- 1,645)
test iter::bench_skip_ref_sum ... bench: 478,207 ns/iter (+/- 119,252)
test iter::bench_skip_sum ... bench: 315,614 ns/iter (+/- 3,054)
test iter::bench_skip_while_chain_ref_sum ... bench: 2,486,370 ns/iter (+/- 4,845)
test iter::bench_skip_while_chain_sum ... bench: 633,915 ns/iter (+/- 5,892)
test iter::bench_skip_while_ref_sum ... bench: 666,926 ns/iter (+/- 804)
test iter::bench_skip_while_sum ... bench: 444,405 ns/iter (+/- 571)
2017-09-25 22:53:08 -05:00
|
|
|
}
|
|
|
|
|
|
|
|
bench_sums! {
|
|
|
|
bench_filter_chain_sum,
|
|
|
|
bench_filter_chain_ref_sum,
|
2019-02-27 04:46:37 -06:00
|
|
|
(0i64..1000000).chain(0..1000000).filter(|x| x % 3 == 0)
|
Add more custom folding to `core::iter` adaptors
Many of the iterator adaptors will perform faster folds if they forward
to their inner iterator's folds, especially for inner types like `Chain`
which are optimized too. The following types are newly specialized:
| Type | `fold` | `rfold` |
| ----------- | ------ | ------- |
| `Enumerate` | ✓ | ✓ |
| `Filter` | ✓ | ✓ |
| `FilterMap` | ✓ | ✓ |
| `FlatMap` | exists | ✓ |
| `Fuse` | ✓ | ✓ |
| `Inspect` | ✓ | ✓ |
| `Peekable` | ✓ | N/A¹ |
| `Skip` | ✓ | N/A² |
| `SkipWhile` | ✓ | N/A¹ |
¹ not a `DoubleEndedIterator`
² `Skip::next_back` doesn't pull skipped items at all, but this couldn't
be avoided if `Skip::rfold` were to call its inner iterator's `rfold`.
Benchmarks
----------
In the following results, plain `_sum` computes the sum of a million
integers -- note that `sum()` is implemented with `fold()`. The
`_ref_sum` variants do the same on a `by_ref()` iterator, which is
limited to calling `next()` one by one, without specialized `fold`.
The `chain` variants perform the same tests on two iterators chained
together, to show a greater benefit of forwarding `fold` internally.
test iter::bench_enumerate_chain_ref_sum ... bench: 2,216,264 ns/iter (+/- 29,228)
test iter::bench_enumerate_chain_sum ... bench: 922,380 ns/iter (+/- 2,676)
test iter::bench_enumerate_ref_sum ... bench: 476,094 ns/iter (+/- 7,110)
test iter::bench_enumerate_sum ... bench: 476,438 ns/iter (+/- 3,334)
test iter::bench_filter_chain_ref_sum ... bench: 2,266,095 ns/iter (+/- 6,051)
test iter::bench_filter_chain_sum ... bench: 745,594 ns/iter (+/- 2,013)
test iter::bench_filter_ref_sum ... bench: 889,696 ns/iter (+/- 1,188)
test iter::bench_filter_sum ... bench: 667,325 ns/iter (+/- 1,894)
test iter::bench_filter_map_chain_ref_sum ... bench: 2,259,195 ns/iter (+/- 353,440)
test iter::bench_filter_map_chain_sum ... bench: 1,223,280 ns/iter (+/- 1,972)
test iter::bench_filter_map_ref_sum ... bench: 611,607 ns/iter (+/- 2,507)
test iter::bench_filter_map_sum ... bench: 611,610 ns/iter (+/- 472)
test iter::bench_fuse_chain_ref_sum ... bench: 2,246,106 ns/iter (+/- 22,395)
test iter::bench_fuse_chain_sum ... bench: 634,887 ns/iter (+/- 1,341)
test iter::bench_fuse_ref_sum ... bench: 444,816 ns/iter (+/- 1,748)
test iter::bench_fuse_sum ... bench: 316,954 ns/iter (+/- 2,616)
test iter::bench_inspect_chain_ref_sum ... bench: 2,245,431 ns/iter (+/- 21,371)
test iter::bench_inspect_chain_sum ... bench: 631,645 ns/iter (+/- 4,928)
test iter::bench_inspect_ref_sum ... bench: 317,437 ns/iter (+/- 702)
test iter::bench_inspect_sum ... bench: 315,942 ns/iter (+/- 4,320)
test iter::bench_peekable_chain_ref_sum ... bench: 2,243,585 ns/iter (+/- 12,186)
test iter::bench_peekable_chain_sum ... bench: 634,848 ns/iter (+/- 1,712)
test iter::bench_peekable_ref_sum ... bench: 444,808 ns/iter (+/- 480)
test iter::bench_peekable_sum ... bench: 317,133 ns/iter (+/- 3,309)
test iter::bench_skip_chain_ref_sum ... bench: 1,778,734 ns/iter (+/- 2,198)
test iter::bench_skip_chain_sum ... bench: 761,850 ns/iter (+/- 1,645)
test iter::bench_skip_ref_sum ... bench: 478,207 ns/iter (+/- 119,252)
test iter::bench_skip_sum ... bench: 315,614 ns/iter (+/- 3,054)
test iter::bench_skip_while_chain_ref_sum ... bench: 2,486,370 ns/iter (+/- 4,845)
test iter::bench_skip_while_chain_sum ... bench: 633,915 ns/iter (+/- 5,892)
test iter::bench_skip_while_ref_sum ... bench: 666,926 ns/iter (+/- 804)
test iter::bench_skip_while_sum ... bench: 444,405 ns/iter (+/- 571)
2017-09-25 22:53:08 -05:00
|
|
|
}
|
|
|
|
|
|
|
|
bench_sums! {
|
|
|
|
bench_filter_map_sum,
|
|
|
|
bench_filter_map_ref_sum,
|
|
|
|
(0i64..1000000).filter_map(|x| x.checked_mul(x))
|
|
|
|
}
|
|
|
|
|
|
|
|
bench_sums! {
|
|
|
|
bench_filter_map_chain_sum,
|
|
|
|
bench_filter_map_chain_ref_sum,
|
|
|
|
(0i64..1000000).chain(0..1000000).filter_map(|x| x.checked_mul(x))
|
|
|
|
}
|
|
|
|
|
|
|
|
bench_sums! {
|
|
|
|
bench_fuse_sum,
|
|
|
|
bench_fuse_ref_sum,
|
|
|
|
(0i64..1000000).fuse()
|
|
|
|
}
|
|
|
|
|
|
|
|
bench_sums! {
|
|
|
|
bench_fuse_chain_sum,
|
|
|
|
bench_fuse_chain_ref_sum,
|
|
|
|
(0i64..1000000).chain(0..1000000).fuse()
|
|
|
|
}
|
|
|
|
|
|
|
|
bench_sums! {
|
|
|
|
bench_inspect_sum,
|
|
|
|
bench_inspect_ref_sum,
|
|
|
|
(0i64..1000000).inspect(|_| {})
|
|
|
|
}
|
|
|
|
|
|
|
|
bench_sums! {
|
|
|
|
bench_inspect_chain_sum,
|
|
|
|
bench_inspect_chain_ref_sum,
|
|
|
|
(0i64..1000000).chain(0..1000000).inspect(|_| {})
|
|
|
|
}
|
|
|
|
|
|
|
|
bench_sums! {
|
|
|
|
bench_peekable_sum,
|
|
|
|
bench_peekable_ref_sum,
|
|
|
|
(0i64..1000000).peekable()
|
|
|
|
}
|
|
|
|
|
|
|
|
bench_sums! {
|
|
|
|
bench_peekable_chain_sum,
|
|
|
|
bench_peekable_chain_ref_sum,
|
|
|
|
(0i64..1000000).chain(0..1000000).peekable()
|
|
|
|
}
|
|
|
|
|
|
|
|
bench_sums! {
|
|
|
|
bench_skip_sum,
|
|
|
|
bench_skip_ref_sum,
|
|
|
|
(0i64..1000000).skip(1000)
|
|
|
|
}
|
|
|
|
|
|
|
|
bench_sums! {
|
|
|
|
bench_skip_chain_sum,
|
|
|
|
bench_skip_chain_ref_sum,
|
|
|
|
(0i64..1000000).chain(0..1000000).skip(1000)
|
|
|
|
}
|
|
|
|
|
|
|
|
bench_sums! {
|
|
|
|
bench_skip_while_sum,
|
|
|
|
bench_skip_while_ref_sum,
|
|
|
|
(0i64..1000000).skip_while(|&x| x < 1000)
|
|
|
|
}
|
|
|
|
|
|
|
|
bench_sums! {
|
|
|
|
bench_skip_while_chain_sum,
|
|
|
|
bench_skip_while_chain_ref_sum,
|
|
|
|
(0i64..1000000).chain(0..1000000).skip_while(|&x| x < 1000)
|
Customize `<FlatMap as Iterator>::fold`
`FlatMap` can use internal iteration for its `fold`, which shows a
performance advantage in the new benchmarks:
test iter::bench_flat_map_chain_ref_sum ... bench: 4,354,111 ns/iter (+/- 108,871)
test iter::bench_flat_map_chain_sum ... bench: 468,167 ns/iter (+/- 2,274)
test iter::bench_flat_map_ref_sum ... bench: 449,616 ns/iter (+/- 6,257)
test iter::bench_flat_map_sum ... bench: 348,010 ns/iter (+/- 1,227)
... where the "ref" benches are using `by_ref()` that isn't optimized.
So this change shows a decent advantage on its own, but much more when
combined with a `chain` iterator that also optimizes `fold`.
2017-09-14 15:51:32 -05:00
|
|
|
}
|
2017-10-23 00:47:27 -05:00
|
|
|
|
|
|
|
bench_sums! {
|
|
|
|
bench_take_while_chain_sum,
|
|
|
|
bench_take_while_chain_ref_sum,
|
|
|
|
(0i64..1000000).chain(1000000..).take_while(|&x| x < 1111111)
|
|
|
|
}
|
2018-03-01 03:57:25 -06:00
|
|
|
|
2018-12-08 05:09:44 -06:00
|
|
|
bench_sums! {
|
|
|
|
bench_cycle_take_sum,
|
|
|
|
bench_cycle_take_ref_sum,
|
|
|
|
(0i64..10000).cycle().take(1000000)
|
|
|
|
}
|
|
|
|
|
2018-03-01 03:57:25 -06:00
|
|
|
// Checks whether Skip<Zip<A,B>> is as fast as Zip<Skip<A>, Skip<B>>, from
|
|
|
|
// https://users.rust-lang.org/t/performance-difference-between-iterator-zip-and-skip-order/15743
|
|
|
|
#[bench]
|
|
|
|
fn bench_zip_then_skip(b: &mut Bencher) {
|
|
|
|
let v: Vec<_> = (0..100_000).collect();
|
|
|
|
let t: Vec<_> = (0..100_000).collect();
|
|
|
|
|
|
|
|
b.iter(|| {
|
|
|
|
let s = v.iter().zip(t.iter()).skip(10000)
|
|
|
|
.take_while(|t| *t.0 < 10100)
|
|
|
|
.map(|(a, b)| *a + *b)
|
|
|
|
.sum::<u64>();
|
|
|
|
assert_eq!(s, 2009900);
|
|
|
|
});
|
|
|
|
}
|
|
|
|
#[bench]
|
|
|
|
fn bench_skip_then_zip(b: &mut Bencher) {
|
|
|
|
let v: Vec<_> = (0..100_000).collect();
|
|
|
|
let t: Vec<_> = (0..100_000).collect();
|
|
|
|
|
|
|
|
b.iter(|| {
|
|
|
|
let s = v.iter().skip(10000).zip(t.iter().skip(10000))
|
|
|
|
.take_while(|t| *t.0 < 10100)
|
|
|
|
.map(|(a, b)| *a + *b)
|
|
|
|
.sum::<u64>();
|
|
|
|
assert_eq!(s, 2009900);
|
|
|
|
});
|
|
|
|
}
|
2019-02-27 04:44:30 -06:00
|
|
|
|
|
|
|
#[bench]
|
|
|
|
fn bench_filter_count(b: &mut Bencher) {
|
|
|
|
b.iter(|| {
|
|
|
|
(0i64..1000000).map(black_box).filter(|x| x % 3 == 0).count()
|
|
|
|
})
|
|
|
|
}
|
|
|
|
|
|
|
|
#[bench]
|
|
|
|
fn bench_filter_ref_count(b: &mut Bencher) {
|
|
|
|
b.iter(|| {
|
|
|
|
(0i64..1000000).map(black_box).by_ref().filter(|x| x % 3 == 0).count()
|
|
|
|
})
|
|
|
|
}
|
|
|
|
|
|
|
|
#[bench]
|
|
|
|
fn bench_filter_chain_count(b: &mut Bencher) {
|
|
|
|
b.iter(|| {
|
|
|
|
(0i64..1000000).chain(0..1000000).map(black_box).filter(|x| x % 3 == 0).count()
|
|
|
|
})
|
|
|
|
}
|
|
|
|
|
|
|
|
#[bench]
|
|
|
|
fn bench_filter_chain_ref_count(b: &mut Bencher) {
|
|
|
|
b.iter(|| {
|
|
|
|
(0i64..1000000).chain(0..1000000).map(black_box).by_ref().filter(|x| x % 3 == 0).count()
|
|
|
|
})
|
2019-02-27 06:22:18 -06:00
|
|
|
}
|