Julia Wiki

Julia (programming language)

From Wikipedia, the free encyclopedia
Jump to navigationJump to search
Julia
Julia prog language.svg
ParadigmMulti-paradigmmultiple dispatch (core), proceduralfunctionalmetamultistaged[1]
Designed byJeff Bezanson, Alan EdelmanStefan KarpinskiViral B. Shah
DeveloperJeff Bezanson, Stefan KarpinskiViral B. Shah, and other contributors[2][3]
First appeared2012; 7 years ago[4]
Stable release
1.2.0[5] / 20 August 2019; 6 days ago
Preview release
1.3.0-rc1[6] / 18 August 2019; 8 days ago / 1.4.0-DEV with daily updates
Typing disciplineDynamicnominativeparametricoptional
Implementation languageJulia, CC++SchemeLLVM[7]
PlatformTier 1: x86-64IA-32CUDA
Tier 2 and 3: ARMv8 and PowerPC
OSLinuxmacOSWindows andFreeBSD
LicenseMIT (core),[2] GPL v2;[7][8] a makefile option omits GPL libraries[9]
Filename extensions.jl
WebsiteJuliaLang.org
Influenced by
Julia is a high-level programming language designed for high-performance numerical analysis and computational science.[13][14][15][16]Distinctive aspects of Julia's design include a type system with parametric polymorphism and types in a fully dynamic programming language and multiple dispatch as its core programming paradigm. It allows concurrentparallel and distributed computing, and direct calling of C and Fortran libraries without glue code. A just-in-time compiler that is referred to as "just-ahead-of-time"[17] in the Julia community is used.
Julia is garbage-collected,[18] uses eager evaluation, and includes efficient libraries for floating-point calculations, linear algebrarandom number generation, and regular expression matching. Many libraries are available, including some (e.g., for fast Fourier transforms) that were previously bundled with Julia and are now separate.[19]
Tools available for Julia include IDEs; with integrated tools, e.g. a linter,[20] debugger,[21] and the Rebugger.jl package "supports repeated-execution debugging"[a] and more.[23]

History[edit]

Work on Julia was started in 2009, by Jeff Bezanson, Stefan KarpinskiViral B. Shah, and Alan Edelman, who set out to create a free language that was both high-level and fast. On 14 February 2012 the team launched a website with a blog post explaining the language's mission.[24] In an interview with InfoWorld in April 2012, Karpinski said of the name "Julia": "There's no good reason, really. It just seemed like a pretty name."[25] Bezanson said he chose the name on the recommendation of a friend.[26]
Since the 2012 launch, the Julia community has grown, with over 9,000,000 downloads from as of June 2019 (and is used at more than 1,500 universities),[27][28][29] The Official Julia Docker images, at Docker Hub, have seen over 4,000,000 downloads as of January 2019.[30][31] The JuliaCon[32] academic conference for Julia users and developers has been held annually since 2014.
Version 0.3 was released in August 2014, version 0.4 in October 2015, and version 0.5 in October 2016.[33] Julia 0.6 was released in June 2017,[34] and was the stable release version until 8 August 2018. Both Julia 0.7 (a useful release for testing packages, and for knowing how to upgrade them for 1.0[35]) and version 1.0 were released on 8 August 2018. Work on Julia 0.7 was a "huge undertaking" (e.g., because of "entirely new optimizer"), and some changes were made to the syntax (with the syntax now stable, and same for 1.x and 0.7) and semantics; the iteration interface was simplified.[36]
The release candidate for Julia 1.0 (Julia 1.0.0-rc1) was released on 7 August 2018, and the final version a day later. Julia 1.1 was released on 21 January 2019 with, e.g., a new "exception stack" language feature. Bugfix releases are expected roughly monthly, for Julia 1.1.x and 1.0.x (1.0.x currently has long-term support; for at least a year) and Julia 1.0.1, up to 1.0.4 have followed that schedule (no such bugfix releases in the pipeline for 0.7-release). Julia 1.2 was released on 20 August 2019 (and with it Julia 1.1.x release are no longer maintained , only 1.0.x and 1.2.x), and it has e.g. some built-in support for web browsers (for testing if running in JavaScript VM).[37]
Most packages that work in Julia 1.0.x also work in 1.1.x or newer, enabled by the forward compatible syntax guarantee. The major exception is, for interacting with non-Julia code, the JavaCall.jl package (however calling other languages, e.g. R language works, with the package for R fixed[38]) to call Java, Scala etc. So to use those languages with Julia, and e.g. JDBC.jl or Apache Spark (through Spark.jl), users can choose to stay with the LTS version of Julia for now,[39] as a milestone is set for a fix in Julia 1.4 (however there's already a workaround for Julia-1.3.0-rc1[40]).

Notable uses[edit]

Julia has attracted some high-profile users, from investment manager BlackRock, which uses it for time-series analytics, to the British insurer Aviva, which uses it for risk calculations. In 2015, the Federal Reserve Bank of New York used Julia to make models of the US economy, noting that the language made model estimation "about 10 times faster" than its previous MATLAB implementation. Julia's co-founders established Julia Computing in 2015 to provide paid support, training, and consulting services to clients, though Julia itself remains free to use. At the 2017 JuliaCon[41] conference, Jeffrey Regier, Keno Fischer and others announced[42] that the Celeste project[43] used Julia to achieve "peak performance of 1.54 petaFLOPS using 1.3 million threads"[44] on 9300 Knights Landing (KNL) nodes of the Cori (Cray XC40) supercomputer (the 5th fastest in the world at the time; by November 2017 was 8th fastest). Julia thus joins C, C++, and Fortran as high-level languages in which petaFLOPS computations have been achieved.
Three of the Julia co-creators are the recipients of the 2019 James H. Wilkinson Prize for Numerical Software (awarded every four years) "for the creation of Julia, an innovative environment for the creation of high-performance tools that enable the analysis and solution of computational science problems."[45]

Sponsors[edit]

Julia has received contributions from 800 developers worldwide.[46] Dr. Jeremy Kepner at MIT Lincoln Laboratory was the founding sponsor of the Julia project in its early days. In addition, funds from the Gordon and Betty Moore Foundation, the Alfred P. Sloan FoundationIntel, and agencies such as NSFDARPANIHNASA, and FAA have been essential to the development of Julia.[47]

Julia Computing[edit]

Julia Computing, Inc. was founded in 2015 by Viral B. Shah, Deepak Vinchhi, Alan Edelman, Jeff Bezanson, Stefan Karpinski and Keno Fischer.[48]
In June 2017, Julia Computing raised $4.6M in seed funding from General Catalyst and Founder Collective.[49]

Language features[edit]

Though designed for numerical computing, Julia is a general-purpose programming language[50]. It is also useful for low-level systems programming,[51] as a specification language,[52] and for web programming: both for server web use[53][54] and for web client programming.[55][56]
According to the official website, the main features of the language are:
  • Multiple dispatch: providing ability to define function behavior across many combinations of argument types
  • Dynamic type system: types for documentation, optimization, and dispatch
  • Good performance, approaching that of statically-typed languages like C
  • A built-in package manager
  • Lisp-like macros and other metaprogramming facilities
  • Call Python functions: use the PyCall package[b]
  • Call C functions directly: no wrappers or special APIs
  • Powerful shell-like abilities to manage other processes
  • Designed for parallel and distributed computing
  • Coroutines: lightweight green threading
  • User-defined types are as fast and compact as built-ins
  • Automatic generation of efficient, specialized code for different argument types
  • Elegant and extensible conversions and promotions for numeric and other types
  • Efficient support for Unicode, including but not limited to UTF-8
Multiple dispatch (also termed multimethods in Lisp) is a generalization of single dispatch – the polymorphic mechanism used in common object-oriented programming (OOP) languages – that uses inheritance. In Julia, all concrete types are subtypes of abstract types, directly or indirectly subtypes of the Any type, which is the top of the type hierarchy. Concrete types can not themselves be subtyped the way they can in other languages; composition is used instead (see also inheritance vs subtyping).
Julia draws significant inspiration from various dialects of Lisp, including Scheme and Common Lisp, and it shares many features with Dylan, also a multiple-dispatch-oriented dynamic language (which features an ALGOL-like free-form infix syntax rather than a Lisp-like prefix syntax, while in Julia "everything"[60] is an expression), and with Fortress, another numerical programming language (which features multiple dispatch and a sophisticated parametric type system). While Common Lisp Object System (CLOS) adds multiple dispatch to Common Lisp, not all functions are generic functions.
In Julia, Dylan, and Fortress extensibility is the default, and the system's built-in functions are all generic and extensible. In Dylan, multiple dispatch is as fundamental as it is in Julia: all user-defined functions and even basic built-in operations like + are generic. Dylan's type system, however, does not fully support parametric types, which are more typical of the ML lineage of languages. By default, CLOS does not allow for dispatch on Common Lisp's parametric types; such extended dispatch semantics can only be added as an extension through the CLOS Metaobject Protocol. By convergent design, Fortress also features multiple dispatch on parametric types; unlike Julia, however, Fortress is statically rather than dynamically typed, with separate compiling and executing phases. The language features are summarized in the following table:
LanguageType systemGeneric functionsParametric types
JuliaDynamicDefaultYes
Common LispDynamicOpt-inYes (but no dispatch)
DylanDynamicDefaultPartial (no dispatch)
FortressStaticDefaultYes
By default, the Julia runtime must be pre-installed as user-provided source code is run. Alternatively, a standalone executable that needs no Julia source code can be built with ApplicationBuilder.jl[61] and PackageCompiler.jl.[62]
Julia's syntactic macros (used for metaprogramming), like Lisp macros, are more powerful and different from text-substitution macros used in the preprocessor of some other languages such as C, because they work at the level of abstract syntax trees (ASTs). Julia's macro system is hygienic, but also supports deliberate capture when desired (like for anaphoric macros) using the esc construct.

Interaction[edit]

The Julia official distribution includes an interactive session shell, called Julia's read–eval–print loop (REPL), which can be used to experiment and test code quickly.[63] The following fragment represents a sample session example where strings are concatenated automatically by println:[64]
julia> p(x) = 2x^2 + 1; f(x, y) = 1 + 2p(x)y
julia> println("Hello world!", " I'm on cloud ", f(0, 4), " as Julia supports recognizable syntax!")
Hello world! I'm on cloud 9 as Julia supports recognizable syntax!
The REPL gives user access to the system shell and to help mode, by pressing ; or ? after the prompt (preceding each command), respectively. It also keeps the history of commands, including between sessions.[65] Code that can be tested inside the Julia's interactive section or saved into a file with a .jl extension and run from the command line by typing:[60]
 $ julia <filename>
Julia is supported by Jupyter, an online interactive "notebooks" environment.[66]

Use with other languages[edit]

Julia is in practice interoperable with many languages. Julia's ccall keyword is used to call C-exported or Fortran shared library functions individually.
Julia has Unicode 12.0 support (and latest 12.1.0 support, which adds only one letter, in Julia 1.3[67]), with UTF-8 used for strings (by default) and for Julia source code, meaning also allowing as an option common math symbols for many operators, such as ∈ for the in operator.
Julia has packages supporting markup languages such as HTML (and also for HTTP), XMLJSON and BSON, and for databases and web use in general.

Implementation[edit]

Julia's core is implemented in Julia and C, together with C++ for the LLVM dependency. The parsing and code-lowering are implemented in FemtoLisp, a Scheme dialect.[68] The LLVM compiler infrastructure project is used as the back end for generation of 64-bit or 32-bit optimized machine code depending on the platform Julia runs on. With some exceptions (e.g., PCRE), the standard library is implemented in Julia itself. The most notable aspect of Julia's implementation is its speed, which is often within a factor of two relative to fully optimized C code (and thus often an order of magnitude faster than Python or R).[69][70][71] Development of Julia began in 2009 and an open-source version was publicized in February 2012.[4][72]

Current and future platforms[edit]

While Julia uses JIT[73] (MCJIT[74] from LLVM) – Julia generates native machine code directly, before a function is first run (not bytecodes that are run on a virtual machine (VM) or translated as the bytecode is running, as with, e.g., Java; the JVM or Dalvik in Android).
Julia has four support tiers,[75] and currently supports all x86-64 processors, that are 64-bit (and is more optimized for the latest generations) and all IA-32 ("x86") processors except for decades old ones, i.e., in 32-bit mode ("i686", excepting CPUs from the pre-Pentium 4-era); and supports more in lower tiers, e.g., tier 2: "fully supports ARMv8 (AArch64) processors, and supports ARMv7 and ARMv6 (AArch32) with some caveats."[76] CUDA (i.e. "Nvidia PTX") has tier 1 support, with the help of an external package.
At least some platforms may need to be compiled from source code (e.g., the original Raspberry Pi), with options changed, while the download page has otherwise executables (and the source) available. Julia has been "successfully built" on several ARM platforms, up to, e.g., "ARMv8 Data Center & Cloud Processors", such as Cavium ThunderX (first ARM with 48 cores). ARM v7 (32-bit) has tier 2 support and binaries (first to get after x86), while ARM v8 (64-bit) and PTX (64-bit) (meaning Nvidia's CUDA on GPUs) has "External" support. PowerPC (64-bit) has tier 3 support "may or may not build". Julia is now supported in Raspbian[77] while support is better for newer, e.g., ARMv7 Pis; the Julia support is promoted by the Raspberry Pi Foundation.[78] Web browsers do not support Julia directly, but a subset of Julia expressions can be translated into JavaScript using JSExpr.jl[55] and a port of Julia to WebAssembly is under development by an external team.[56]

See also[edit]

Notes[edit]

  1. ^ [With Rebugger.jl] you can:
    • test different modifications to the code or arguments as many times as you want; you are never forced to exit “debug mode” and save your file
    • run the same chosen block of code repeatedly (perhaps trying out different ways of fixing a bug) without needing to repeat any of the “setup” work that might have been necessary to get to some deeply nested method in the original call stack.[22]
  2. ^ For calling the newer Python 3 (the older default to call Python 2, is also still supported)[57][58] and calling in the other direction, from Python to Julia, is also supported with pyjulia.[59]

No comments:

Post a Comment