paradis

paradis is currently at an early, experimental stage. Test coverage is deliberately poor in order to make it easier to iterate on the overall design. Community feedback is very welcome!

paradis makes it easier to implement non-trivial parallel algorithms that require access to a subset of indices into data structures that are structurally similar to multidimensional arrays. It does so by providing abstractions at incrementally higher levels:

1. A low-level, unsafe abstraction for unsynchronized access to independent records of a collection.
2. Higher-level abstractions built on top of the unsafe base layer that allow many parallel access patterns to be expressed in safe code, or with a minimum of unsafe code.

The low-level abstractions are provided by the very lightweight paradis-core crate. Library authors are encouraged to depend only on this crate in order to expose their data structures for parallel access.

Please check out the documentation for more information about how to use paradis.

Examples

The examples given here are provided just to give you a taste of the API. Please check out the documentation for more context.

Safe parallel iteration with index lists

The following example shows how paradis can be used to safely iterate over mutable elements located at arbitrary indices in a slice, in parallel.

use paradis::index::{IndexList, narrow_access};
use paradis::rayon::create_par_iter;
use rayon::iter::ParallelIterator;

let mut data = vec![0, 1, 2, 3, 4, 5, 6, 7, 8, 9];
let indices = vec![4, 7, 1].check_unique().expect("Indices are unique");
let access = narrow_access(data.as_mut_slice(), &indices)
    .expect("Indices are in bounds of the data structure");
create_par_iter(access).for_each(|x_i| *x_i = 0);

assert_eq!(data, vec![0, 0, 2, 3, 0, 5, 6, 0, 8, 9]);

Structured index lists

For some problems, the indices are structured. In this case, we may be able to avoid runtime checks for uniqueness, and instead prove uniqueness by structured construction, using index combinators. The example below shows how structured uniqueness allows us to mutate the superdiagonal of a matrix.

use nalgebra::dmatrix;
use paradis::index::{IndexList, narrow_access};
use paradis::rayon::create_par_iter;
use rayon::iter::ParallelIterator;

// Access implementation omitted
use paradis_demo::DMatrixParAccessMut;

let mut matrix = dmatrix![1, 1, 1, 1, 1;
                          1, 1, 1, 1, 1;
                          1, 1, 1, 1, 1];

// Superdiagonal indices are [(0, 1), (1, 2), (2, 3)]
let superdiagonal_indices = (0 .. 3).index_zip(1 .. 4);
let access = DMatrixParAccessMut::from_matrix_mut(&mut matrix);
let superdiagonal_access = narrow_access(access, &superdiagonal_indices)
    .expect("Indices are in bounds");

create_par_iter(superdiagonal_access).for_each(|x_ij| *x_ij = 0);

assert_eq!(matrix,
           dmatrix![1, 0, 1, 1, 1;
                    1, 1, 0, 1, 1;
                    1, 1, 1, 0, 1]);

Low-level unsafe parallel access

The higher-level features of paradis are built on top of its low-level abstractions for parallel access. The example below shows how we may use careful unsynchronized access to mutate even and odd parts of a slice in different threads.

use paradis_core::{BoundedParAccess, IntoParAccess};
use std::thread::scope;

let mut data = vec![0; 100];
let n = data.len();
let access = data.into_par_access();

scope(|s| {
    s.spawn(|| {
        // The first thread touches elements at even indices
        for i in (0 .. n).step_by(2) {
            unsafe { *access.get_unsync(i) = 1; }
        }
    });

    s.spawn(|| {
        // The second thread touches elements at odd indices
        for i in (1 .. n).step_by(2) {
            unsafe { *access.get_unsync(i) = 2; }
        }
    });
})

Contributing

paradis is open source, and contribution is welcome. There are several ways you can contribute:

1. by trying paradis out for your application and reporting on your experience in the forum.
2. by filing issues, like bugs, ideas for improvement or concerns about the design.
3. by fixing bugs, improving documentation, or contributing new features as part of a pull request. Please note that not at all new features will necessarily be accepted. Before investing a great deal of time on new functionality, please file an issue to see if the feature is likely to be accepted.

Keep in mind that paradis is not developed professionally. Although I, @Andlon, have every intention of following up on issues and PRs, life has a tendency to get in the way at times, sometimes for extended periods of time.

License

paradis is distributed under the terms of either the MIT license or the Apache License (Version 2.0), at your option. See LICENSE-APACHE and LICENSE-MIT for details.

By contributing intellectual property to this repository, you agree to license your contribution under the same terms.