homepage, bench improvement and doc fix #317

dhardy · 2018-03-20T15:45:34Z

No description provided.

pitdicker · 2018-03-20T16:24:49Z

benches/generators.rs

                }
+                black_box(accum);


I had on my wish-list to investigate the benchmark 'problems'. So nice to see 😄. I also noticed some change if you set b.bytes before b.iter. Is that something you can confirm?

Never tried that.

Something similar might be possible for the fill_bytes stuff, but it's probably less significant (already this has little impact on 64-bit outputs).

It made a difference, but not any more with the accumulator.

pitdicker · 2018-03-20T16:25:34Z

I would like to take a closer look, that will probably be tomorrow.

dhardy · 2018-03-20T17:25:52Z

Added a bit more doc. I hope it's not too scary 😱

At this point maybe it's not worth mentioning the <R: Rng>(rng: R) variant at all actually; I tried implementing it but had some issues; see also #287 (comment)

pitdicker · 2018-03-21T08:20:11Z

Your change really helps with the new parallel codegen. The same would be necessary for the distributions code. I gave it a try in https://github.com/pitdicker/rand/commits/doc_bench. Do you want to take that commit, or maybe do the same in some cleaner way?

Then there is only one test left that is badly affected by parallel codegen: misc_sample_indices_10_of_1k. I don't see an easy solution there yet.

pitdicker · 2018-03-21T08:23:21Z

Cargo.toml

@@ -6,7 +6,7 @@ license = "MIT/Apache-2.0"
 readme = "README.md"
 repository = "https://github.com/rust-lang-nursery/rand"
 documentation = "https://docs.rs/rand"
-homepage = "https://github.com/rust-lang-nursery/rand"
+homepage = "https://crates.io/crates/rand"


I don't like either 😄. Every time I look up a crate on crates.io is surprisingly unhelpful, and I quickly click through to the repository or docs.rs.

But this seems like a good change.

pitdicker · 2018-03-21T08:25:31Z

rand-core/Cargo.toml

@@ -4,9 +4,9 @@ version = "0.1.0-pre.0"
 authors = ["The Rust Project Developers"]
 license = "MIT/Apache-2.0"
 readme = "README.md"
-repository = "https://github.com/rust-lang-nursery/rand"
+repository = "https://github.com/rust-lang-nursery/rand/tree/master/rand-core"


I don't think this is right, as you can't do git checkout https://github.com/rust-lang-nursery/rand/tree/master/rand-core. This should just be the base repository, and cargo figures things out.

pitdicker · 2018-03-21T08:25:54Z

rand-core/Cargo.toml

@@ -4,9 +4,9 @@ version = "0.1.0-pre.0"
 authors = ["The Rust Project Developers"]
 license = "MIT/Apache-2.0"
 readme = "README.md"
-repository = "https://github.com/rust-lang-nursery/rand"
+repository = "https://github.com/rust-lang-nursery/rand/tree/master/rand-core"
 documentation = "https://docs.rs/rand"


Does this line need changing?
(documentation = "https://docs.rs/rand")

pitdicker · 2018-03-21T08:28:41Z

src/lib.rs

+///   `(&mut rng).gen()`, (3) in some cases `mut rng: R` is required, and (4)
+///   since `R` may not be a reference (it has no reborrow requirement) and no
+///   `Copy` bound, there is no choice but to move or double-reference uses of
+///   `rng: R` within `foo`. Both 2 and 3 are probably Rust bugs, but 4 isn't


This is a sentence of 12 lines. Do you happen to be a lawyer? 😄

pitdicker · 2018-03-21T08:36:59Z

src/lib.rs

+/// - `fn foo<R: Rng>(rng: &mut R)` will not match type-erased `Rng`s; in order
+///   to support dynamic dispatch use either
+///   `fn foo<R: Rng + ?Sized>(rng: &mut R)` or `fn foo<R: Rng>(rng: R)`
+///   (but see next point about the latter).


Added a bit more doc. I hope it's not too scary 😱

You give the technical reasons, and that is supposed to be a plus for programmers. But maybe it is good to help users a little. Maybe add something like:

Tip if you want to use Rng in trait bounds:
The basic pattern is fn foo<R: Rng + ?Sized>(rng: &mut R).

This way implementations can use the methods from both RngCore and Rng, support dynamic dispatch (thanks to ?Sized), and ...

Technical details:
...

pitdicker · 2018-03-21T10:17:16Z

benches/distributions.rs

                let mut accum: $ty = 0;
                for _ in 0..::RAND_BENCH_N {
                    accum = accum.wrapping_add(rng.gen_range($low, high));
-                    high += fake_increment;
+                    high += $incr;  // force recalculation of range each time


How about high = high.wrapping_add(1) & std::%ty::MAX, so it also works for i8? (Not thought deeply about it to be honest)

I think that would still wrap, and end up less than low? Current version seems to work anyway.

dhardy · 2018-03-21T10:35:13Z

I removed the generic usage doc (still working on that stuff). Rest is okay to merge?

pitdicker · 2018-03-21T10:36:19Z

Yes. What did you think of #317 (comment)?

dhardy · 2018-03-21T11:26:40Z

Didn't you see my reply?

pitdicker · 2018-03-21T11:34:00Z

Sorry. I was comparing the benchmarks with and without parallel codegen, and checking the generated assembly on master, your branch, and my suggestion with a mask.

But without the mask, all the range code is completely optimized out, at least for i8. I don't completely get why. I pushed a commit the the same branch as before.

dhardy · 2018-03-21T11:40:32Z

Ah, i8 is signed; took me a while to understand why binary and worked. 👍

homepage, bench improvement and doc fix

dhardy requested a review from pitdicker March 20, 2018 15:45

pitdicker reviewed Mar 20, 2018

View reviewed changes

pitdicker reviewed Mar 21, 2018

View reviewed changes

dhardy force-pushed the doc branch from a30e9ab to 9b6c387 Compare March 21, 2018 10:33

dhardy and others added 5 commits March 21, 2018 12:06

Generator benchmarks: reduce overhead for u32/u64 generation

064749b

Distribution benchmarks: reduce overhead

02e3c0e

Distribution benchmarks: further tweaks; improve range bench perf

e70480d

rand-core: update homepage location and fix relative link

46b383c

Use mask in gen_range_i benchmarks

a76487c

dhardy force-pushed the doc branch from 00ebb7b to a76487c Compare March 21, 2018 12:06

dhardy added P-high Priority: high D-review Do: needs review labels Mar 21, 2018

pitdicker merged commit 4634912 into rust-random:master Mar 21, 2018

dhardy deleted the doc branch March 23, 2018 17:51

pitdicker added a commit that referenced this pull request Apr 4, 2018

Merge pull request #317 from dhardy/doc

dbc1f1d

homepage, bench improvement and doc fix

pitdicker mentioned this pull request May 12, 2018

Add CONTRIBUTING.md #443

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

homepage, bench improvement and doc fix #317

homepage, bench improvement and doc fix #317

dhardy commented Mar 20, 2018

pitdicker Mar 20, 2018

dhardy Mar 20, 2018

pitdicker Mar 21, 2018

pitdicker commented Mar 20, 2018

dhardy commented Mar 20, 2018

pitdicker commented Mar 21, 2018

pitdicker Mar 21, 2018 •

edited

Loading

pitdicker Mar 21, 2018

pitdicker Mar 21, 2018 •

edited

Loading

pitdicker Mar 21, 2018

pitdicker Mar 21, 2018

pitdicker Mar 21, 2018

dhardy Mar 21, 2018

dhardy commented Mar 21, 2018

pitdicker commented Mar 21, 2018

dhardy commented Mar 21, 2018

pitdicker commented Mar 21, 2018

dhardy commented Mar 21, 2018

homepage, bench improvement and doc fix #317

homepage, bench improvement and doc fix #317

Conversation

dhardy commented Mar 20, 2018

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

pitdicker commented Mar 20, 2018

dhardy commented Mar 20, 2018

pitdicker commented Mar 21, 2018

pitdicker Mar 21, 2018 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

pitdicker Mar 21, 2018 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

dhardy commented Mar 21, 2018

pitdicker commented Mar 21, 2018

dhardy commented Mar 21, 2018

pitdicker commented Mar 21, 2018

dhardy commented Mar 21, 2018

pitdicker Mar 21, 2018 •

edited

Loading

pitdicker Mar 21, 2018 •

edited

Loading