In all versions of Ruby up to and including Ruby 1.8.x, there was no access to “native” threads (that is, threads handled by the operating system). In effect, Ruby 1.8 threads exist inside the closed world of a Ruby program, with multiple threads each being allocated time, using a procedure called time-slicing, within a single process. Ruby 1.9 (and newer) uses a new interpreter, YARV (Yet Another Ruby Virtual-machine). This allows Ruby 1.9 to make use of native threads, albeit with some limitations that I’ll explain shortly.

In principle, native threads allow more efficient execution (using preemptive multitasking) whereby the operating system takes care of the execution of threads on one or more processors. Even though Ruby 1.9 uses native threads, it does not perform preemptive multitasking. For reasons of compatibility with existing Ruby programs, Ruby 1.9 native threads work in a similar fashion to Ruby 1.8 non-native (or green) threads. In other words, although Ruby 1.9 may in fact run a native thread, it is the Ruby virtual machine, rather than the operating system, that schedules the execution of threads. This means Ruby threads sacrifice efficiency; however, they do at least benefit from portability: Threads written on one operating system will also run on a different operating system.