luke5542 / languagecomparison Goto Github PK

It would be useful to add in several other popular languages to the benchmarks (not limited to):

• PHP
• Swift (once the Linux runtime has been released)
• F#
• Haskell

Possibly missing shebang.

This makes it easier for examples like Rust, Go, and D where they can benefit from using build tools such as cargo.

Will also make it cleaner for adding examples of different algorithms in the future.

Requiring everyone to download 50 separate runtimes in order to run the tests is a bit funky. Using Docker, you could download a standard execution environment for each language, and execute the benchmark within, with no performance overhead.

Only some of the languages, like Python, have this correctly working. It needs to be fixed for all of the other languages as well.

Should probably make sure that the examples validate all the inputs (more useful with timing now done inside the examples).

Change each example to measure time taken from within the code to minimise the impact of overheads, especially for setup in scripting examples.

We should probably work to include LDC and GDC compilers for D, in a similar manner to Clang vs GCC that we already have for C/C++.

Right now it's pretty difficult to distinguish between the results over multiple runs of the benchmarks. All the languages at sub-one-second get mashed together and it's not really possible to compare them.