Which languages, compilers, and OSes are supported?¶
The definitive list ships with each build
New language versions and compiler support are added regularly. The authoritative, current list for your build is in the New Project wizard and under Project ▸ Configure Project ▸ Languages. The tables below are a guide.
Languages¶
Understand can analyze a single language or combine several in one project (see mixed-language codebases).
| Language | Dialects / versions supported |
|---|---|
| Ada | Ada 83, 95, 2005, 2012 — separately or combined |
| Assembly | Coldfire 68k, JIPSE MIL-STD-1750A, IBM System/370 |
| C / C++ | K&R and ANSI C; C++ through C++23; Objective-C and Objective-C++; CUDA. Works with most compilers |
| C# | Through C# 12 |
| Fortran | FORTRAN 77, 90, 95, 2003, 2008, 2018 (fixed and free format); Harris and DEC extensions |
| Java | Through Java 21 |
| JOVIAL | JOVIAL73 and JOVIAL3 |
| Pascal / Delphi | Embarcadero Delphi and Turbo Pascal; ISO 7185:1990 (Unextended Pascal) with HP extensions |
| Python | Python 2 and Python 3 |
| Rust | Rust |
| VHDL | VHDL-87, -93, -2002, -2008 |
| Visual Basic | VB / VB.NET, 2002 through 2022 |
| Web | HTML, CSS, JavaScript, TypeScript, PHP, XML |
Missing a compiler extension?
Understand is regularly extended to cover common compiler extensions. If a construct you use isn't parsed, contact support.
Operating systems¶
| OS | Supported versions | Architecture |
|---|---|---|
| Windows | Windows 11 (x64) | 64-bit |
| Linux | RHEL / CentOS 8.6+, Ubuntu 22.04+ (and later) | 64-bit x86_64 only (not ARM Linux) |
| macOS | Monterey (12) and later | Universal — Intel and Apple Silicon |
Older operating systems
If you must run on an older OS, SciTools publishes older builds with earlier OS baselines. Contact support for the right build.
See Install Understand for per-platform install steps and library requirements.
Hardware and space (rough guidance)¶
- Disk: roughly 600 MB – 1.1 GB for the installation, depending on OS and whether a bundled AI model is included.
- Memory: about 1 GB of RAM per million lines of code is a good rule of thumb — it runs on less but more is faster.
- AI features (optional): running a local AI model has its own, heavier requirements (dedicated GPU or a strong CPU, and several GB of free VRAM/RAM). See AI overview and setup. You can install a build without a bundled model if you don't need it — see the kit variants in Download & create an account.