Green thread explained

Green thread should not be confused with virtual thread.

In computer programming, a green thread is a thread that is scheduled by a runtime library or virtual machine (VM) instead of natively by the underlying operating system (OS). Green threads emulate multithreaded environments without relying on any native OS abilities, and they are managed in user space instead of kernel space, enabling them to work in environments that do not have native thread support.^[1]

Etymology

Green threads refers to the name of the original thread library for Java programming language (that was released in version 1.1 and then Green threads were abandoned in version 1.3 to native threads). It was designed by The Green Team at Sun Microsystems.^[2]

History

Green threads were briefly available in Java between 1997 and 2000.

Green threads share a single operating system thread through co-operative concurrency and can therefore not achieve parallelism performance gains like operating system threads. The main benefit of coroutines and green threads is ease of implementation.

Performance

On a multi-core processor, native thread implementations can automatically assign work to multiple processors, whereas green thread implementations normally cannot.^{[1] [3]} Green threads can be started much faster on some VMs. On uniprocessor computers, however, the most efficient model has not yet been clearly determined.

Benchmarks on computers running the Linux kernel version 2.2 (released in 1999) have shown that:^[4]

• Green threads significantly outperform Linux native threads on thread activation and synchronization.
• Linux native threads have slightly better performance on input/output (I/O) and context switching operations.

When a green thread executes a blocking system call, not only is that thread blocked, but all of the threads within the process are blocked.^[5] To avoid that problem, green threads must use asynchronous I/O operations, although the increased complexity on the user side can be reduced if the virtual machine implementing the green threads spawns specific I/O processes (hidden to the user) for each I/O operation.

There are also mechanisms which allow use of native threads and reduce the overhead of thread activation and synchronization:

• Thread pools reduce the cost of spawning a new thread by reusing a limited number of threads.^[6]
• Languages which use virtual machines and native threads can use escape analysis to avoid synchronizing blocks of code when unneeded.^[7]

Green threads in the Java Virtual Machine

In Java 1.1, green threads were the only threading model used by the Java virtual machine (JVM),^[8] at least on Solaris. As green threads have some limitations compared to native threads, subsequent Java versions dropped them in favor of native threads.^{[9] [10]}

An exception to this is the Squawk virtual machine, which is a mixture between an operating system for low-power devices and a Java virtual machine. It uses green threads to minimize the use of native code, and to support migrating its isolates.

Kilim^{[11] [12]} and Quasar^{[13] [14]} are open-source projects which implement green threads on later versions of the JVM by modifying the Java bytecode produced by the Java compiler (Quasar also supports Kotlin and Clojure).

Green threads in other languages

There are some other programming languages that implement equivalents of green threads instead of native threads. Examples:

• Chicken Scheme uses lightweight user-level threads based on first-class continuations^[15]
• Common Lisp^[16]
• CPython natively supports asyncio since Version 3.4, alternative implementations exist like greenlet, eventlet and gevent, PyPy^[17]
• Crystal offers fibers^[18]
• D offers fibers, used for asynchronous I/O^[19]
• Dyalog APL terms them threads^[20]
• Erlang^[21]
• Go implements so called goroutines^[22]
• Haskell
• Julia uses green threads for its Tasks.
• Limbo^[23]
• Lua uses coroutines for concurrency. Lua 5.2 also offers true C coroutine semantics through the functions lua_yieldk, lua_callk, and lua_pcallk. The CoCo extension allows true C coroutine semantics for Lua 5.1.
• Nim provides asynchronous I/O and coroutines
• OCaml, since version 5.0, supports green threads through the Domainslib.Task module
• occam, which prefers the term process instead of thread due to its origins in communicating sequential processes
• Perl supports green threads through coroutines
• PHP supports green threads through fibers and coroutines
• Ruby before version 1.9^[24]
• Racket (native threads are also available through Places^[25])
• Rust supports system threads natively^[26] and supports asynchronous I/O through third-party libraries like Tokio
• SML/NJ's implementation of Concurrent ML
• Smalltalk (most dialects: Squeak, VisualWorks, GNU Smalltalk, etc.)
• Stackless Python supports either preemptive multitasking or cooperative multitasking through microthreads (termed tasklets).^[27]
• Tcl has coroutines and an event loop^[28]

The Erlang virtual machine has what might be called green processes – they are like operating system processes (they do not share state like threads do) but are implemented within the Erlang Run Time System (erts). These are sometimes termed green threads, but have significant differences from standard green threads.

In the case of GHC Haskell, a context switch occurs at the first allocation after a configurable timeout. GHC threads are also potentially run on one or more OS threads during their lifetime (there is a many-to-many relationship between GHC threads and OS threads), allowing for parallelism on symmetric multiprocessing machines, while not creating more costly OS threads than needed to run on the available number of cores.

Most Smalltalk virtual machines do not count evaluation steps; however, the VM can still preempt the executing thread on external signals (such as expiring timers, or I/O becoming available). Usually round-robin scheduling is used so that a high-priority process that wakes up regularly will effectively implement time-sharing preemption: [[(Delay forMilliseconds: 50) wait] repeat ] forkAt: Processor highIOPriorityOther implementations, e.g., QKS Smalltalk, are always time-sharing. Unlike most green thread implementations, QKS also supports preventing priority inversion.

Differences to virtual threads in the Java Virtual Machine

Virtual threads were introduced as a preview feature in Java 19^[29] and stabilized in Java 21.^[30] Important differences between virtual threads and green threads are:

• Virtual threads coexist with existing (non-virtual) platform threads and thread pools.
• Virtual threads protect their abstraction:
  • Unlike with green threads, sleeping on a virtual thread does not block the underlying carrier thread.
  • Working with thread-local variables is deemphasized, and scoped values are suggested as a more lightweight replacement.^[31]
• Virtual threads can be cheaply suspended and resumed, making use of JVM support for the special jdk.internal.vm.Continuation class.
• Virtual threads handle blocking calls by transparently unmounting from the carrier thread where possible, otherwise compensating by increasing the number of platform threads.

See also

• Async/await
• Light-weight process
• Coroutine
• Java virtual machine
• Global interpreter lock
• Fiber (computer science)
• GNU Portable Threads
• Protothreads

External links

• "Four for the ages", JavaWorld article about Green threads
• Green threads on Java threads FAQ

Notes and References

1. Web site: Sintes . Tony . 2001-04-13 . mdy . Four for the ages . . 2020-07-14 . Green threads, the threads provided by the JVM, run at the user level, meaning that the JVM creates and schedules the threads itself. Therefore, the operating system kernel doesn't create or schedule them. Instead, the underlying OS sees the JVM only as one thread. Green threads prove inefficient for a number of reasons. Foremost, green threads cannot take advantage of a multiprocessor system(...) Thus, the JVM threads are bound to run within that single JVM thread that runs inside a single processor. . 2020-07-15 . https://web.archive.org/web/20200715050100/https://www.infoworld.com/article/2077383/four-for-the-ages.html . bot: unknown .
2. Web site: https://web.archive.org/web/20080530073139/http://java.sun.com/features/1998/05/birthday.html . 2008-05-30 . Java Technology: The Early Years . java.sun.com . 2014-12-22.
3. Web site: What is the difference between "green" threads and "native" threads?. jguru.com. 2000-09-06. 2009-06-01. On multi-CPU machines, native threads can run more than one thread simultaneously by assigning different threads to different CPUs. Green threads run on only one CPU..
4. Comparative performance evaluation of Java threads for embedded applications: Linux Thread vs. Green Thread . 10.1.1.8.9238.
5. Book: Stallings, William . Operating Systems, Internal and Design Principles . 2008 . Prentice Hall . New Jersey . 9780136006329 . 171.
6. Web site: Concurrency in JRuby. Engine Yard. Sieger. Nick. 2011-07-22. 2013-01-26. For systems with large volumes of email, this naive approach may not work well. Native threads carry a bigger initialization cost and memory overhead than green threads, so JRuby normally cannot support more than about 10,000 threads. To work around this, we can use a thread pool.. 2014-01-30. https://web.archive.org/web/20140130094407/http://blog.engineyard.com/2011/concurrency-in-jruby. dead.
7. Web site: Java theory and practice: Synchronization optimizations in Mustang. IBM. Goetz . Brian. 2005-10-18. 2013-01-26.
8. Web site: Java Threads in the Solaris Environment – Earlier Releases. Oracle Corporation. 2013-01-26. As a result, several problems arose: Java applications could not interoperate with existing MT applications in the Solaris environment, Java threads could not run in parallel on multiprocessors, An MT Java application could not harness true OS concurrency for faster applications on either uniprocessors or multiprocessors. To substantially increase application performance, the green threads library was replaced with native Solaris threads for Java on the Solaris 2.6 platform; this is carried forward on the Solaris 7 and Solaris 8 platforms..
9. Web site: Threads: Green or Native. SCO Group. 2013-01-26. The performance benefit from using native threads on an MP machine can be dramatic. For example, using an artificial benchmark where Java threads are doing processing independent of each other, there can be a three-fold overall speed improvement on a 4-CPU MP machine..
10. Web site: Threads: Green or Native. codestyle.org. 2013-01-26. https://web.archive.org/web/20130116024929/http://www.codestyle.org/java/faq-Threads.shtml#greenthread. 2013-01-16. dead. There is a significant processing overhead for the JVM to keep track of thread states and swap between them, so green thread mode has been deprecated and removed from more recent Java implementations..
11. Web site: kilim . GitHub . 2016-06-09.
12. Web site: Kilim . www.malhar.net . 2016-06-09.
13. Web site: Quasar Code on GitHub. GitHub.
14. Web site: Parallel Universe . 6 December 2015 . 22 December 2015 . https://web.archive.org/web/20151222173703/http://www.paralleluniverse.co/quasar/ . dead .
15. Web site: Chicken Scheme . 5 November 2017.
16. Web site: thezerobit/green-threads. GitHub . 2016-04-08.
17. Web site: Application-level Stackless features – PyPy 4.0.0 documentation. 6 December 2015.
18. Web site: Concurrency: GitBook . crystal-lang.org . 2018-04-03.
19. Web site: Fibers - Dlang Tour . 2022-05-02 . tour.dlang.org.
20. Web site: Threads: Overview . Dyalog APL 17.0 Help . 2018-12-14 . "A thread is a strand of execution in the APL workspace.".
21. 1010485913393254401. joeerl. Erlang processes are emulated in the Erlang VM, like Green threads - we like them since this simplifies many proble… . 23 June 2018.
22. Web site: Go and Dogma . research!rsc . 2017-01-14 . "for example both Go and Haskell need some kind of “green threads”, so there are more shared runtime challenges than you might expect.".
23. Web site: The Limbo Programming Language. www.vitanuova.com. 2019-04-01.
24. Web site: Multithreading in the MRI Ruby Interpreter BugFactory . 2024-06-18.
25. Web site: Racket Places. 2011-10-13. Places enable the development of parallel programs that take advantage of machines with multiple processors, cores, or hardware threads. A place is a parallel task that is effectively a separate instance of the Racket virtual machine..
26. Web site: Using Threads to Run Code Simultaneously - The Rust Programming Language. 2021-09-24. doc.rust-lang.org.
27. Web site: Stackless.com: About Stackless . A round robin scheduler is built in. It can be used to schedule tasklets either cooperatively or preemptively. . 2008-08-27 . dead . https://web.archive.org/web/20120227125313/http://zope.stackless.com/about/sdocument_view . 2012-02-27.
28. Web site: Tcl event loop. 6 December 2015.
29. Web site: JEP 425: Virtual Threads (Preview). 2024-01-25.
30. Web site: JEP 444: Virtual Threads. 2024-01-25.
31. Web site: JEP 464: Scoped Values (Second Preview). 2024-01-25.