A "shootout" in the programming world is often the comparison of a wide spectrum of languages or interpreters. In the past, there have been shootouts spanning tens of languages and variations of languages, testing similar programs. Both program length and execution time were measured, and there were some unexpected and interesting results. Of course, there are plenty of tables with exact times, and much more easily digestible graphs to look at. It's all great fun, root for your favorites, see which ones are fastest and which ones fall flat. And marvel at how much memory the Java programs use.
But of particular interest is the latest Ruby shootout. Among many others, it tests the latest Ruby 1.9.2 RC. Have things improved any further since 1.9.1? In short, yes, but not by far (at least as far as performance is concerned). The clear winners in this shootout are 1.9.1 and 1.9.2, beating out all the others in most tests. This shootout also compares memory utilization, with 1.9.x being the clear winner yet again. jRuby, no surprise since it runs on Java, uses the most. And another interesting aspect is tested, Linux vs. Windows performance. This isn't saying that much about the platforms themselves, since the Ruby implementations on Windows are often undermaintained, but Linux is significantly faster for Ruby. I'd like to see some Linux vs. OSX benchmarks though, as those are the two "serious" Ruby platforms.
Overall, these shootouts are great fun and a little useful, but they're not real-life benchmarks. The results don't always extrapolate to real life performance in the same manner. They do give a pretty good indication though.
>> In the past, there have been shootouts spanning tens of languages and variations of languages, testing similar programs. <<
Not – "In the past".
The Computer Language Benchmarks Game is still updated:
- here are the summary measurements for 3 different Ruby implementations on x86 Ubuntu.
- here’s the JRuby :: Ruby 1.9.1 comparison on x64 Ubuntu.