benchmarksgame-rs

Further update: I now have a Rust program that is 100% safe code and is also faster than any of the unsafe programs (including the C and C++ entries): https://github.com/TeXitoi/benchmarksgame-rs/pull/55

As an update, this pull request now removes most of the unsafe code: https://github.com/TeXitoi/benchmarksgame-rs/pull/46

Thanks! I like Rust for what it does in advancing the state of the art of the languages, but I also like how this example demonstrates how hard it is to avoid "unsafe" constructs and remain competitive. https://github.com/TeXitoi/benchmarksgame-rs/blob/master/src...

> afaict comparing Rust program #4 [5.30 secs] to Rust program #5 [9.14 secs] is all about differences between std::collections::HashMap and that experimental hash table. There was also a small contribution (~6%) of working on bytes rather than strings according to the original PR: https://github.com/TeXitoi/benchmarksgame-rs/pull/39

In other Rust/benchmarksgame news, I just submitted a simple fix to the Rust program for "reverse-complement" that makes it faster than the fastest C++ program, on my computer. The old version was spending 2/3 of its time just reading the input into memory, because it wasn't allocating a large enough buffer up front. https://github.com/TeXitoi/benchmarksgame-rs/pull/44 I'm also working on some additional changes that make it even faster than the C version (again, on my computer) by improving how it divides work across CPUs: https://github.com/TeXitoi/benchmarksgame-rs/pull/46 These improved Rust programs have not yet been added to the benchmarksgame site. Previous entries are ranked at: http://benchmarksgame.alioth.debian.org/u64q/performance.php... Minor improvements to the Rust programs for "mandelbrot" and "binary-trees" are also awa

I don't know the history here, but https://github.com/TeXitoi/benchmarksgame-rs/blob/master/src... https://github.com/TeXitoi/benchmarksgame-rs/issues/34 (though some things have changed since these comments...)

Here are a bunch of them: https://github.com/TeXitoi/benchmarksgame-rs/tree/master/src Here's a grep for unsafe: https://github.com/TeXitoi/benchmarksgame-rs/search?utf8=%E2... 3/14 benchmarks use unsafe. pidigits just uses unsafe to bind to GMP. reverse-complement is weird and I'm not sure what's going on there at first glance. spectralnorm looks like it's just using it to divide up an array in parallel; in the past couple of months there have a been several Rust libraries to enable this from safe code, so that may be able to be removed now.

My experience with compiler development is that the incentives all align toward making the "default RNG that people go to" as fast as possible to the exclusion of all else, including the quality of the generated numbers. That's because people frequently write benchmarks gated on the speed of random number generation, and those benchmarks usually don't care about the quality of the random numbers. Sometimes those benchmarks get popular, which encourages a race to the bottom in RNG quality. (Similar incentives exist for hash functions, by the way.) Popular RNG-bound benchmarks: * "fasta" from the Benchmarks Game [1] (thankfully, this one mandates use of its own RNG, though it's vulnerable to fast-forwarding as demonstrated in [2]) * "Perlin noise" [3] (almost entirely RNG performance bound) * From SunSpider (this one being particularly rele