jmbatto/juliabench repository overview

⁠Julia 1.12 & PDI Development Environment (Debian Trixie)

A robust, source-compiled development environment designed for high-performance coupling between Julia and C using PDI. This image is built on Debian Trixie (Testing) and includes critical patches for stability with modern toolchains (GCC 13+).
Docker file on github⁠

⁠🚀 Key Features

• Julia 1.12.4 (Source Compiled): Built from source to ensure compatibility with Debian Trixie's system libraries.

  • Stability Fixes: Compiled with -gdwarf-4 and noexecstack to resolve critical libunwind segfaults caused by GCC 13 defaults.
  • System Integration: Configured to use internal unwinding while avoiding conflicts with system LLVM.

• PDI 1.10.0: Installed with full Python support.

  • Compatibility: Patched to support Python 3.12+ (which removed distutils), ensuring the pycall plugin compiles correctly.

• Headless GTK4 Support: Pre-configured with Xvfb, dbus-x11, and libgl1 to support graphical plotting libraries (ProfileView, Gtk4.jl) in a headless container environment.

• Clean Environment: LD_LIBRARY_PATH is strictly optimized to prevent "DLL Hell" between Julia Artifacts and system libraries (GLib/GObject).

⁠📦 Installed Components

• OS: Debian Trixie (Testing)
• Julia: v1.12.4
• PDI: v1.10.0 (with HDF5, Python, and PyCall plugins)
• Python: 3.12+ (with numpy and setuptools)
• Dev Tools: build-essential, hdf5-tools, cmake, git, gdb, valgrind, neowofetch.
• Julia Packages: HDF5, DataFrames, Gtk4, ProfileView, PProf.

⁠🛠 Usage

⁠Basic Interactive Shell

docker run -it jmbatto/juliabench

⁠Important

To use Gtk4 or ProfileView without a physical display, you must wrap the command with xvfb-run :

xvfb-run --auto-servernum --server-args="-screen 0 1920x1080x24 -nolisten tcp" \
    julia --project=. -e 'using ProfileView; ...'