/julia

The Julia Language: A fresh approach to technical computing.

Primary LanguageJuliaOtherNOASSERTION

Julia Logo

Linux, OSX: Build Status

Windows: Build status

Code Coverage: Coverage Status codecov.io

## The Julia Language

Julia is a high-level, high-performance dynamic language for technical computing. The main homepage for Julia can be found at julialang.org. This is the GitHub repository of Julia source code, including instructions for compiling and installing Julia, below.

## Resources

The mailing list for developer discussion is http://groups.google.com/group/julia-dev/. All are welcome, but the volume of messages is higher, and the discussions tend to be more esoteric. New developers may find the notes in CONTRIBUTING helpful to start contributing to the Julia codebase.

External Resources

## Currently Supported Platforms
  • GNU/Linux
  • Darwin/OS X
  • FreeBSD
  • Windows

All systems are supported with both x86/64 (64-bit) and x86 (32-bit) architectures. Experimental and early support for ARM is available too.

## Source Download and Compilation

First, acquire the source code by cloning the git repository:

git clone git://github.com/JuliaLang/julia.git

(If you are behind a firewall, you may need to use the https protocol instead of the git protocol:

git config --global url."https://".insteadOf git://

Be sure to also configure your system to use the appropriate proxy settings, e.g. by setting the https_proxy and http_proxy variables.)

By default you will be building the latest unstable version of Julia. However, most users should use the most recent stable version of Julia, which is currently the 0.4 series of releases. You can get this version by changing to the Julia directory and running

git checkout release-0.4

Now run make to build the julia executable. To perform a parallel build, use make -j N and supply the maximum number of concurrent processes. When compiled the first time, it will automatically download and build its external dependencies. This takes a while, but only has to be done once. If the defaults in the build do not work for you, and you need to set specific make parameters, you can save them in Make.user. The build will automatically check for the existence of Make.user and use it if it exists. Building Julia requires 1.5GiB of disk space and approximately 700MiB of virtual memory.

For builds of julia starting with 0.5.0-dev, you can create out-of-tree builds of Julia by specifying make O=<build-directory> configure on the command line. This will create a directory mirror, with all of the necessary Makefiles to build Julia, in the specified directory. These builds will share the source files in Julia and deps/srccache. Each out-of-tree build directory can have its own Make.user file to override the global Make.user file in the top-level folder.

If you need to build Julia in an environment that does not allow access to the outside world, use make -C deps getall to download all the necessary files. Then, copy the julia directory over to the target environment and build with make.

Note: the build process fail badly if any of the build directory's parent directories have spaces or other shell meta-characters such as $ or : in their names (this is due to a limitation in GNU make).

Once it is built, you can run the julia executable using its full path in the directory created above (the julia directory), or, to run it from anywhere, either

  • add a soft link to the julia executable in the julia directory to /usr/local/bin (or any suitable directory already in your path), or

  • add the julia directory to your executable path for this shell session (in bash: export PATH="$(pwd):$PATH" ; in csh or tcsh: set path= ( $path $cwd ) ), or

  • add the julia directory to your executable path permanently (e.g. in .bash_profile), or

  • write prefix=/path/to/install/folder into Make.user and then run make install. If there is a version of Julia already installed in this folder, you should delete it before running make install.

Now you should be able to run Julia like this:

julia

If everything works correctly, you will see a Julia banner and an interactive prompt into which you can enter expressions for evaluation. (Errors related to libraries might be caused by old, incompatible libraries sitting around in your PATH. In that case, try moving the julia directory earlier in the PATH).

Your first test of Julia should be to determine whether your build is working properly. From the UNIX/Windows command prompt inside the julia source directory, type make testall. You should see output that lists a series of tests being run; if they complete without error, you should be in good shape to start using Julia.

You can read about getting started in the manual.

If you are building a Julia package for distribution on Linux, OS X, or Windows, take a look at the detailed notes in DISTRIBUTING.md.

Updating an existing source tree

If you have previously downloaded julia using git clone, you can update the existing source tree using git pull rather than starting anew:

cd julia
git pull && make

Assuming that you had made no changes to the source tree that will conflict with upstream updates, these commands will trigger a build to update to the latest version.

General troubleshooting

  1. Over time, the base library may accumulate enough changes such that the bootstrapping process in building the system image will fail. If this happens, the build may fail with an error like

     *** This error is usually fixed by running 'make clean'. If the error persists, try 'make cleanall' ***

    As described, running make clean && make is usually sufficient. Occasionally, the stronger cleanup done by make cleanall is needed.

  2. New versions of external dependencies may be introduced which may occasionally cause conflicts with existing builds of older versions.

    a. Special make targets exist to help wipe the existing build of a dependency. For example, make -C deps clean-llvm will clean out the existing build of llvm so that llvm will be rebuilt from the downloaded source distribution the next time make is called. make -C deps distclean-llvm is a stronger wipe which will also delete the downloaded source distribution, ensuring that a fresh copy of the source distribution will be downloaded and that any new patches will be applied the next time make is called.

    b. To delete existing binaries of julia and all its dependencies, delete the ./usr directory in the source tree.

  3. If you've updated OS X recently, be sure to run xcode-select --install to update the command line tools. Otherwise, you could run into errors for missing headers and libraries, such as ld: library not found for -lcrt1.10.6.o.

  4. If you've moved the source directory, you might get errors such as CMake Error: The current CMakeCache.txt directory ... is different than the directory ... where CMakeCache.txt was created., in which case you may delete the offending dependency under deps

  5. In extreme cases, you may wish to reset the source tree to a pristine state. The following git commands may be helpful:

     git reset --hard #Forcibly remove any changes to any files under version control
     git clean -x -f -d #Forcibly remove any file or directory not under version control

    To avoid losing work, make sure you know what these commands do before you run them. git will not be able to undo these changes!

## Uninstalling Julia

Julia does not install anything outside the directory it was cloned into. Julia can be completely uninstalled by deleting this directory. Julia packages are installed in ~/.julia by default, and can be uninstalled by deleting ~/.julia.

## Platform-Specific Build Notes

General

  • GCC version 4.7 or later is required to build Julia.
  • To use external shared libraries not in the system library search path, set USE_SYSTEM_XXX=1 and LDFLAGS=-Wl,-rpath,/path/to/dir/contains/libXXX.so in Make.user.
    • Instead of setting LDFLAGS, putting the library directory into the environment variable LD_LIBRARY_PATH (at both compile and run time) also works.
  • See also the external dependencies.

Architecture Customization

Julia can be built for a non-generic architecture by configuring the ARCH Makefile variable. See the appropriate section of Make.inc for additional customization options, such as MARCH and JULIA_CPU_TARGET.

For example, to build for Pentium 4, set MARCH=pentium4 and install the necessary system libraries for linking. On Ubuntu, these may include lib32gfortran3 (also manually call ln -s /usr/lib32/libgfortran3.so.0 /usr/lib32/libgfortran3.so) and lib32gcc1, lib32stdc++6, among others.

You can also set MARCH=native for a maximum-performance build customized for the current machine CPU.

Ubuntu

The julia-deps PPA contains updated packages for Julia dependencies if you want to use system libraries instead of having them downloaded and built during the build process. See System Provided Libraries.

For a fast and easy current installation, the before_install section of travis.yml is a great resource. Note that those instructions are for Ubuntu 12.04, and for later versions you may need to install newer versions of dependencies, such as libunwind8-dev instead of libunwind7-dev.

RHEL/CentOS 6

On RHEL/CentOS 6 systems, the default compiler (gcc 4.4) is too old to build Julia.

Install or contact your systems administrator to install a more recent version of gcc. The Scientific Linux Developer Toolset works well.

Linux Build Troubleshooting

Problem Possible Solution
OpenBLAS build failure Set one of the following build options in Make.user and build again:
  • OPENBLAS_TARGET_ARCH=BARCELONA (AMD CPUs) or OPENBLAS_TARGET_ARCH=NEHALEM (Intel CPUs)
      Set OPENBLAS_DYNAMIC_ARCH = 0 to disable compiling multiple architectures in a single binary.
  • OPENBLAS_NO_AVX2 = 1 disables AVX2 instructions, allowing OpenBLAS to compile with OPENBLAS_DYNAMIC_ARCH = 1 using old versions of binutils
  • USE_SYSTEM_BLAS=1 uses the system provided libblas
    • Set LIBBLAS=-lopenblas and LIBBLASNAME=libopenblas to force the use of the system provided OpenBLAS when multiple BLAS versions are installed.

If you get an error that looks like ../kernel/x86_64/dgemm_kernel_4x4_haswell.S:1709: Error: no such instruction: `vpermpd $ 0xb1,%ymm0,%ymm0', then you need to set OPENBLAS_DYNAMIC_ARCH = 0 or OPENBLAS_NO_AVX2 = 1, or you need a newer version of binutils (2.18 or newer). (Issue #7653)

Illegal Instruction error Check if your CPU supports AVX while your OS does not (e.g. through virtualization, as described in this issue).

OS X

You need to have the current Xcode command line utilities installed: run xcode-select --install in the terminal. You will need to rerun this terminal command after each OS X update, otherwise you may run into errors involving missing libraries or headers. You will also need a 64-bit gfortran to compile Julia dependencies. The gfortran-4.7 (and newer) compilers in Brew, Fink, and MacPorts work for building Julia.

Clang is now used by default to build Julia on OS X 10.7 and above. On OS X 10.6, the Julia build will automatically use gcc. On current systems, we recommend that you install the command line tools as described above. Older systems do not have a separate command line tools package from Apple, and will require a full Xcode install. On these, you will need at least Xcode 4.3.3. In Xcode prior to v5.0, you can alternatively go to Preferences -> Downloads and select the Command Line Utilities. These steps will ensure that clang v3.1 is installed, which is the minimum version of clang required to build Julia.

If you have set LD_LIBRARY_PATH or DYLD_LIBRARY_PATH in your .bashrc or equivalent, Julia may be unable to find various libraries that come bundled with it. These environment variables need to be unset for Julia to work.

If you see build failures in OpenBLAS or if you prefer to experiment, you can use the Apple provided BLAS in vecLib by building with USE_SYSTEM_BLAS=1. Julia does not use the Apple provided LAPACK, as it is too old.

When building Julia, or its dependencies, libraries installed by third party package managers can redirect the compiler to use an incompatible version of the software it is looking for. One example of this happening is when a piece of software called the "linker" gives an error involving "Undefined symbols." If that happens, you can usually figure out what software package is causing the error from the names in the error text. This sort of error can be bypassed by, temporarily, uninstalling the offending package. If the offending package cannot be uninstalled by itself, it may be possible to just uninstall the development headers (for example: a package ending in "-dev" in Fink).

FreeBSD

On FreeBSD Release 9.0, install the gcc47, git, and gmake packages/ports, and compile Julia with the command:

$ gmake FC=gfortran47

You must use the gmake command on FreeBSD instead of make.

Windows

In order to build Julia on Windows, see README.windows.

Vagrant

Julia can be developed in an isolated Vagrant environment. See the Vagrant README for details.

## Required Build Tools and External Libraries

Building Julia requires that the following software be installed:

  • GNU make — building dependencies.
  • gcc & g++ (>= 4.7) or Clang (>= 3.1, Xcode 4.3.3 on OS X) — compiling and linking C, C++
  • gfortran — compiling and linking Fortran libraries
  • git — version control and package management (version 1.7.3+ required)
  • perl — preprocessing of header files of libraries.
  • wget, curl, or fetch (FreeBSD) — to automatically download external libraries.
  • m4 — needed to build GMP.
  • patch — for modifying source code.
  • cmake — needed to build libgit2.
  • openssl — needed for HTTPS support in libgit2 on Linux, install via apt-get install libssl-dev or yum install openssl-devel.

Julia uses the following external libraries, which are automatically downloaded (or in a few cases, included in the Julia source repository) and then compiled from source the first time you run make:

  • LLVM (3.7) — compiler infrastructure.
  • FemtoLisp — packaged with Julia source, and used to implement the compiler front-end.
  • libuv — portable, high-performance event-based I/O library
  • OpenLibm — portable libm library containing elementary math functions.
  • OpenSpecFun (>= 0.4) — library containing Bessel and error functions of complex arguments.
  • DSFMT — fast Mersenne Twister pseudorandom number generator library.
  • OpenBLAS — fast, open, and maintained basic linear algebra subprograms (BLAS) library, based on Kazushige Goto's famous GotoBLAS.
  • LAPACK (>= 3.5) — library of linear algebra routines for solving systems of simultaneous linear equations, least-squares solutions of linear systems of equations, eigenvalue problems, and singular value problems.
  • MKL (optional) – OpenBLAS and LAPACK may be replaced by Intel's MKL library.
  • AMOS — subroutines for computing Bessel and Airy functions.
  • SuiteSparse (>= 4.1) — library of linear algebra routines for sparse matrices.
  • ARPACK — collection of subroutines designed to solve large, sparse eigenvalue problems.
  • FFTW (>= 3.3) — library for computing fast Fourier transforms very quickly and efficiently.
  • PCRE (>= 8.31) — Perl-compatible regular expressions library.
  • GMP (>= 5.0) — GNU multiple precision arithmetic library, needed for BigInt support.
  • MPFR (>= 3.0) — GNU multiple precision floating point library, needed for arbitrary precision floating point (BigFloat) support.
  • libgit2 (>= 0.21) — Git linkable library, used by Julia's package manager
  • utf8proc (>= 1.3) — a library for processing UTF-8 encoded Unicode strings
  • libosxunwind — clone of libunwind, a library that determines the call-chain of a program
  • Rmath-julia — library for commonly used statistical functions from the R project.

For a longer overview of Julia's dependencies, see these slides.

### System Provided Libraries

If you already have one or more of these packages installed on your system, you can prevent Julia from compiling duplicates of these libraries by passing USE_SYSTEM_...=1 to make or adding the line to Make.user. The complete list of possible flags can be found in Make.inc.

Please be aware that this procedure is not officially supported, as it introduces additional variability into the installation and versioning of the dependencies, and is recommended only for system package maintainers. Unexpected compile errors may result, as the build system will do no further checking to ensure the proper packages are installed.

SuiteSparse is a special case, since it is typically only installed as a static library, while USE_SYSTEM_SUITESPARSE=1 requires that it is a shared library. Running the script contrib/repackage_system_suitesparse4.make will copy your static system SuiteSparse installation into the shared library format required by Julia. make USE_SYSTEM_SUITESPARSE=1 will then use the SuiteSparse that has been copied into Julia's directory, but will not build a new SuiteSparse library from scratch.

Intel compilers and Math Kernel Library (MKL)

To build Julia using the Intel compilers (icc, icpc, ifort), and link against the MKL BLAS and LAPACK libraries, first make sure you have a recent version of the compiler suite (version 15 or later).

For a 64-bit architecture, the environment should be set up as follows:

# bash
source /path/to/intel/bin/compilervars.sh intel64

Add the following to the Make.user file:

USEICC = 1
USEIFC = 1
USE_INTEL_MKL = 1
USE_INTEL_MKL_FFT = 1
USE_INTEL_LIBM = 1

It is highly recommended to start with a fresh clone of the Julia repository.

## Source Code Organization

The Julia source code is organized as follows:

base/          source code for Julia's standard library
contrib/       editor support for Julia source, miscellaneous scripts
deps/          external dependencies
doc/manual     source for the user manual
doc/stdlib     source for standard library function help text
examples/      example Julia programs
src/           source for Julia language core
test/          test suites
test/perf      benchmark suites
ui/            source for various front ends
usr/           binaries and shared libraries loaded by Julia's standard libraries
## Binary Installation

Because of the rapid pace of development at this point, we recommend installing the latest Julia from source, but platform-specific tarballs with pre-compiled binaries are also available for download.

You can either run the julia executable using its full path in the directory created above, or add that directory to your executable path so that you can run the Julia program from anywhere (in the current shell session):

export PATH="$(pwd)/julia:$PATH"

Now you should be able to run Julia like this:

julia

On Windows, double-click usr/bin/julia.exe.

If everything works correctly, you will see a Julia banner and an interactive prompt into which you can enter expressions for evaluation. You can read about getting started in the manual.

The following distributions include julia, but the versions may be out of date due to rapid development:

## Editor and Terminal Setup

Currently, Julia editing mode support is available for Emacs, Vim, Textmate, Sublime Text, Notepad++, and Kate, in contrib/. There is early support for IDEs such as Juno, Eclipse (LiClipse). A notebook interface is available through IJulia, which adds Julia support to Jupyter/IPython. The Sublime-IJulia plugin enables interaction between IJulia and Sublime Text.

In the terminal, Julia makes great use of both control-key and meta-key bindings. To make the meta-key bindings more accessible, many terminal emulator programs (e.g., Terminal, iTerm, xterm, etc.) allow you to use the alt or option key as meta. See the section in the manual on interacting with Julia for more details.