Codee 2024.4 released

We’re pleased to announce Codee 2024.4, the latest release of our automated code inspection platform for Fortran, C, and C++. The new Codee capabilities ensure code correctness, modernization, and optimization by enforcing coding guidelines that elevate code quality, modernize legacy code, increase portability, and enhance performance. Codee accelerates software delivery, empowering teams to meet industry standards and achieve their quality and efficiency goals.

What’s new

• New AutoFixes:
  • PWR069: It automatically adds the `only` keyword to `use` statements to only import what is really needed from a module.
  • PWR072: It splits the variable initialization from the declaration to prevent the implicit ‘save’ behavior.
• Improved AutoFix: the PWR007 AutoFix was updated to handle GNU Fortran intrinsic procedures.
• New Fortran checker: new PWR075 for porting code with GNU extensions to Fortran standards.
• LLVM Toolchain Upgrade: Updated to version 20 (20.0.0) for enhanced performance and compatibility.
• Intel Compiler Compatibility: Added support for Intel ifx compiler flags on both Linux and Windows.
• Fortran native multithreading support: Introduced a “native” option for multithreading AutoFixes, which uses the construct `do concurrent` in Fortran loops instead of OpenMP pragmas.
• Bear Integration: Bundled Bear binary with Codee to allow users to create compile_commands.json files out of the box. Special thanks to the Bear team for maintaining this project.
• Enhanced Source File Dependency Manager: Improved reporting on missing Fortran modules, helping users resolve dependencies and expand code analysis coverage.
• Automatic compile_commands Detection: Codee now autodetects compile_commands.json files in the current working directory.
• Conditional OpenMP Support: Added compatibility for Fortran programs with conditionally compiled OpenMP lines.

We’ll be at SC24 to present this new release. Click here to schedule a meeting and discover how Codee can streamline your HPC projects!

We also presented the results of our recent work, Optimizing the Weather Research and Forecasting Model (WRF) with OpenMP Offload and Codee, a collaboration with Berkeley Lab, NERSC and the U.S. Department of Energy at the WACCPD workshop.

Stay tuned for more updates and stay connected with us through our newsletter.

Happy coding!