Upgrade to Pro
— share decks privately, control downloads, hide ads and more …
Speaker Deck
Features
Speaker Deck
PRO
Sign in
Sign up for free
Search
Search
Rayon (Rust Belt Rust)
Search
nikomatsakis
October 28, 2016
Programming
7
1k
Rayon (Rust Belt Rust)
A talk about Rayon from the Rust Belt Rust conference
nikomatsakis
October 28, 2016
Tweet
Share
More Decks by nikomatsakis
See All by nikomatsakis
Hereditary Harrop Formulas (Papers We Love Boston)
nikomatsakis
2
460
Rust: Systems Programming for All!
nikomatsakis
0
170
CppNow 2017
nikomatsakis
0
190
Rust at Mozilla (part of Mozilla Onboarding)
nikomatsakis
0
160
Guaranteeing Memory Safety and Data-Race Freedom in Rust
nikomatsakis
0
220
Other Decks in Programming
See All in Programming
fieldalignmentから見るGoの構造体
kuro_kurorrr
0
130
Lambda(Python)の リファクタリングが好きなんです
komakichi
4
240
2ヶ月で生産性2倍、お買い物アプリ「カウシェ」4チーム同時改善の取り組み
ike002jp
1
110
プロダクトエンジニアのしごと 〜 受託 × 高難度を乗り越えるOptium開発 〜
algoartis
0
160
Носок на сок
bo0om
0
1.1k
「理解」を重視したAI活用開発
fast_doctor
0
270
ComposeでWebアプリを作る技術
tbsten
0
130
Road to RubyKaigi: Making Tinny Chiptunes with Ruby
makicamel
4
540
Qiita Bash
mercury_dev0517
2
220
Boost Your Performance and Developer Productivity with Jakarta EE 11
ivargrimstad
0
760
Jakarta EE Meets AI
ivargrimstad
0
780
flutter_kaigi_mini_4.pdf
nobu74658
0
140
Featured
See All Featured
Bootstrapping a Software Product
garrettdimon
PRO
307
110k
10 Git Anti Patterns You Should be Aware of
lemiorhan
PRO
656
60k
GraphQLの誤解/rethinking-graphql
sonatard
71
10k
Into the Great Unknown - MozCon
thekraken
38
1.7k
ピンチをチャンスに:未来をつくるプロダクトロードマップ #pmconf2020
aki_iinuma
119
51k
Docker and Python
trallard
44
3.4k
The Cult of Friendly URLs
andyhume
78
6.3k
Building Flexible Design Systems
yeseniaperezcruz
329
39k
The Psychology of Web Performance [Beyond Tellerrand 2023]
tammyeverts
47
2.7k
Testing 201, or: Great Expectations
jmmastey
42
7.5k
The Invisible Side of Design
smashingmag
299
50k
Faster Mobile Websites
deanohume
306
31k
Transcript
Rayon Data Parallelism for Fun and Profit Nicholas Matsakis (nmatsakis
on IRC)
Want to make parallelization easy 2 fn load_images(paths: &[PathBuf]) ->
Vec<Image> { paths.iter() .map(|path| Image::load(path)) .collect() } fn load_images(paths: &[PathBuf]) -> Vec<Image> { paths.par_iter() .map(|path| Image::load(path)) .collect() } For each path… …load an image… …create and return a vector.
Want to make parallelization safe 3 fn load_images(paths: &[PathBuf]) ->
Vec<Image> { let mut pngs = 0; paths.par_iter() .map(|path| { if path.ends_with(“png”) { pngs += 1; } Image::load(path) }) .collect() } Data-race Will not compile
4 http://blog.faraday.io/saved-by-the-compiler-parallelizing-a-loop-with-rust-and-rayon/
5 Parallel Iterators join() threadpool Basically all safe Safe interface
Unsafe impl Unsafe
6 fn load_images(paths: &[PathBuf]) -> Vec<Image> { paths.iter() .map(|path| Image::load(path))
.collect() }
7 fn load_images(paths: &[PathBuf]) -> Vec<Image> { paths.par_iter() .map(|path| Image::load(path))
.collect() }
Not quite that simple… 8 (but almost!) 1. No mutating
shared state (except for atomics, locks). 2. Some combinators are inherently sequential. 3. Some things aren’t implemented yet.
9 fn load_images(paths: &[PathBuf]) -> Vec<Image> { let mut pngs
= 0; paths.par_iter() .map(|path| { if path.ends_with(“png”) { pngs += 1; } Image::load(path) }) .collect() } Data-race Will not compile
10 `c` not shared between iterations! fn increment_all(counts: &mut [u32])
{ for c in counts.iter_mut() { *c += 1; } } fn increment_all(counts: &mut [u32]) { paths.par_iter_mut() .for_each(|c| *c += 1); }
fn load_images(paths: &[PathBuf]) -> Vec<Image> { let pngs = paths.par_iter()
.filter(|p| p.ends_with(“png”)) .map(|_| 1) .sum(); paths.par_iter() .map(|p| Image::load(p)) .collect() } 11
12 But beware: atomics introduce nondeterminism! use std::sync::atomic::{AtomicUsize, Ordering}; fn
load_images(paths: &[PathBuf]) -> Vec<Image> { let pngs = AtomicUsize::new(0); paths.par_iter() .map(|path| { if path.ends_with(“png”) { pngs.fetch_add(1, Ordering::SeqCst); } Image::load(path) }) .collect() }
13 3 2 1 12 0 4 5 1 2
1 3 2 1 0 1 3 4 0 3 6 7 8 vec1 vec2 6 2 6 * sum 8 82 fn dot_product(vec1: &[i32], vec2: &[i32]) -> i32 { vec1.iter() .zip(vec2) .map(|(e1, e2)| e1 * e2) .fold(0, |a, b| a + b) // aka .sum() }
14 fn dot_product(vec1: &[i32], vec2: &[i32]) -> i32 { vec1.par_iter()
.zip(vec2) .map(|(e1, e2)| e1 * e2) .reduce(|| 0, |a, b| a + b) // aka .sum() } 3 2 1 12 0 4 5 1 2 1 3 2 1 0 1 3 4 0 3 6 7 8 vec1 vec2 sum 20 19 43 39 82
15 Parallel iterators: Mostly like normal iterators, but: • closures
cannot mutate shared state • some operations are different For the most part, Rust protects you from surprises.
16 Parallel Iterators join() threadpool
The primitive: join() 17 rayon::join(|| do_something(…), || do_something_else(…)); Meaning: maybe
execute two closures in parallel. Idea: - add `join` wherever parallelism is possible - let the library decide when it is profitable
18 fn load_images(paths: &[PathBuf]) -> Vec<Image> { paths.par_iter() .map(|path| Image::load(path))
.collect() } Image::load(paths[0]) Image::load(paths[1])
Work stealing 19 Cilk: http://supertech.lcs.mit.edu/cilk/ (0..22) Thread A Thread B
(0..15) (15..22) (1..15) (queue) (queue) (0..1) (15..22) (15..18) (18..22) (15..16) (16..18) “stolen” (18..22) “stolen”
20
21 Parallel Iterators join() threadpool Rayon: • Parallelize for fun
and profit • Variety of APIs available • Future directions: • more iterators • integrate SIMD, array ops • integrate persistent trees • factor out threadpool
22 Parallel Iterators join() scope() threadpool
23 the scope `s` task `t1` task `t2` rayon::scope(|s| {
… s.spawn(move |s| { // task t1 }); s.spawn(move |s| { // task t2 }); … });
rayon::scope(|s| { … s.spawn(move |s| { // task t1 s.spawn(move
|s| { // task t2 … }); … }); … }); 24 the scope task t1 task t2
`not_ok` is freed here 25 the scope task t1 let
ok: &[u32]s = &[…]; rayon::scope(|scope| { … let not_ok: &[u32] = &[…]; … scope.spawn(move |scope| { // which variables can t1 use? }); });
26 fn join<A,B>(a: A, b: B) where A: FnOnce() +
Send, B: FnOnce() + Send, { rayon::scope(|scope| { scope.spawn(move |_| a()); scope.spawn(move |_| b()); }); } (Real join avoids heap allocation)
27 struct Tree<T> { value: T, children: Vec<Tree<T>>, } impl<T>
Tree<T> { fn process_all(&mut self) { process_value(&mut self.value); for child in &mut self.children { child.process_all(); } } }
28 impl<T> Tree<T> { fn process_all(&mut self) where T: Send
{ rayon::scope(|scope| { for child in &mut self.children { scope.spawn(move |_| child.process_all()); } process_value(&mut self.value); }); } }
29 impl<T> Tree<T> { fn process_all(&mut self) where T: Send
{ rayon::scope(|scope| { let children = &mut self.children; scope.spawn(move |scope| { for child in &mut children { scope.spawn(move |_| child.process_all()); } }); process_value(&mut self.value); }); } }
30 impl<T: Send> Tree<T> { fn process_all(&mut self) { rayon::scope(|s|
self.process_in(s)); } fn process_in<‘s>(&’s mut self, scope: &Scope<‘s>) { let children = &mut self.children; scope.spawn(move |scope| { for child in &mut children { scope.spawn(move |scope| child.process_in(scope)); } }); process_value(&mut self.value); } }