Skip to content

Commit

Permalink
Improve std::thread::available_parallelism docs
Browse files Browse the repository at this point in the history
  • Loading branch information
yoshuawuyts committed Oct 13, 2021
1 parent 44995f7 commit 21429ed
Showing 1 changed file with 57 additions and 20 deletions.
77 changes: 57 additions & 20 deletions library/std/src/thread/mod.rs
Original file line number Diff line number Diff line change
Expand Up @@ -1427,39 +1427,76 @@ fn _assert_sync_and_send() {
_assert_both::<Thread>();
}

/// Returns the number of hardware threads available to the program.
///
/// This value should be considered only a hint.
///
/// # Platform-specific behavior
///
/// If interpreted as the number of actual hardware threads, it may undercount on
/// Windows systems with more than 64 hardware threads. If interpreted as the
/// available concurrency for that process, it may overcount on Windows systems
/// when limited by a process wide affinity mask or job object limitations, and
/// it may overcount on Linux systems when limited by a process wide affinity
/// mask or affected by cgroups limits.
/// Returns an estimate of the default amount of parallelism a program should use.
///
/// Parallelism is a resource. A given machine provides a certain capacity for
/// parallelism, i.e., a bound on the number of computations it can perform
/// simultaneously. This number often corresponds to the amount of CPUs or
/// computer has, but it may diverge in various cases.
///
/// Host environments such as VMs or container orchestrators may want to
/// restrict the amount of parallelism made available to programs in them. This
/// is often done to limit the potential impact of (unintentionally)
/// resource-intensive programs on other programs running on the same machine.
///
/// # Limitations
///
/// The purpose of this API is to provide an easy and portable way to query
/// the default amount of parallelism the program should use. Among other things it
/// does not expose information on NUMA regions, does not account for
/// differences in (co)processor capabilities, and will not modify the program's
/// global state in order to more accurately query the amount of available
/// parallelism.
///
/// The value returned by this function should be considered a simplified
/// approximation of the actual amount of parallelism available at any given
/// time. To get a more detailed or precise overview of the amount of
/// parallelism available to the program, you may wish to use
/// platform-specific APIs as well. The following platform limitations currently
/// apply to `available_parallelism`:
///
/// On Windows:
/// - It may undercount the amount of parallelism available on systems with more
/// than 64 logical CPUs. However, programs typically need specific support to
/// take advantage of more than 64 logical CPUs, and in the absence of such
/// support, the number returned by this function accurately reflects the
/// number of logical CPUs the program can use by default.
/// - It may overcount the amount of parallelism available on systems limited by
/// process-wide affinity masks, or job object limitations.
///
/// On Linux:
/// - It may overcount the amount of parallelism available when limited by a
/// process-wide affinity mask, or when affected by cgroup limits.
///
/// On all targets:
/// - It may overcount the amount of parallelism available when running in a VM
/// with CPU usage limits (e.g. an overcommitted host).
///
/// # Errors
///
/// This function will return an error in the following situations, but is not
/// limited to just these cases:
/// This function will, but is not limited to, return errors in the following
/// cases:
///
/// - If the number of hardware threads is not known for the target platform.
/// - The process lacks permissions to view the number of hardware threads
/// available.
/// - If the amount of parallelism is not known for the target platform.
/// - If the program lacks permission to query the amount of parallelism made
/// available to it.
///
/// # Examples
///
/// ```
/// # #![allow(dead_code)]
/// #![feature(available_parallelism)]
/// use std::thread;
/// use std::{io, thread};
///
/// let count = thread::available_parallelism().map(|n| n.get()).unwrap_or(1);
/// fn main() -> io::Result<()> {
/// let count = thread::available_parallelism()?.get();
/// assert!(count >= 1_usize);
/// Ok(())
/// }
/// ```
#[doc(alias = "available_concurrency")] // Alias for a previous name we gave this API on unstable.
#[doc(alias = "hardware_concurrency")] // Alias for C++ `std::thread::hardware_concurrency`.
#[doc(alias = "available_concurrency")] // Alias for a name we gave this API on unstable.
#[doc(alias = "num_cpus")] // Alias for a popular ecosystem crate which provides similar functionality.
#[unstable(feature = "available_parallelism", issue = "74479")]
pub fn available_parallelism() -> io::Result<NonZeroUsize> {
imp::available_parallelism()
Expand Down

0 comments on commit 21429ed

Please sign in to comment.