cpu-time - Small crate that provides CPU time measurement

README
Issues 4

CPU Time Measurement Library

A simple and idiomatic interface for measurement CPU time:

let start = ProcessTime::now();
# .. do something ..
let cpu_time: Duration = start.elapsed();
println!(" {:?}");

License

Licensed under either of

Apache License, Version 2.0, (./LICENSE-APACHE or http://www.apache.org/licenses/LICENSE-2.0)
MIT license (./LICENSE-MIT or http://opensource.org/licenses/MIT) at your option.

Contribution

Unless you explicitly state otherwise, any contribution intentionally submitted for inclusion in the work by you, as defined in the Apache-2.0 license, shall be dual licensed as above, without any additional terms or conditions.

1 Open More issues

Closed

Add non-panicking variants of the functions.

While it's true that CLOCK_PROCESS_CPUTIME_ID on Linux has been there for very long, but for other nixes the support is much more recent and was added in 2013 for both OpenBSD and FreeBSD. On OSX clock_gettime is unsupported before Sierra (2016)

I'm not really sure about Windows, but as far as I understand, this function may fail depending on the security/privilege configuration and how the process/thread was created, see this MSDN page

I also ran rustfmt on the parts of the code changed by me. You may consider reformatting the whole tree using rustfmt and adding the style check to the CI.

opened Oct 7, 2019 by marmistrz 12

Closed

monotonicity guarantees

Found by @matklad: monotonicity isn't guaranteed, especially when the process is migrated between CPUs. https://docs.rs/cpu-time/1.0.0/src/cpu_time/clock_gettime.rs.html#38

unsafe { clock_gettime(CLOCK_PROCESS_CPUTIME_ID, &mut time) }
...
    /// Returns the amount of CPU time used from the previous timestamp.
    pub fn duration_since(&self, timestamp: Self) -> Duration {
        self.0 - timestamp.0
    }

From https://linux.die.net/man/3/clock_gettime:

The CLOCK_PROCESS_CPUTIME_ID and CLOCK_THREAD_CPUTIME_ID clocks are realized on many platforms using timers from the CPUs (TSC on i386, AR.ITC on Itanium). These registers may differ between CPUs and as a consequence these clocks may return bogus results if a process is migrated to another CPU.

So seems like this can panic? https://doc.rust-lang.org/stable/src/core/time.rs.html#875-881

opened Feb 3, 2021 by ailisp 3

Closed