AdaCore has used the Embedded World Conference this week to announce the availability of its GNAT Pro Safety-Critical product for ARM Cortex micro-controllers. This bareboard product provides an Ada development environment oriented towards systems that are safety-critical in nature, or those that have stringent memory constraints.
White PapersMore >>
- High Performance Computing in Finance: Best Practices Revealed
- Client Windows Migration: Expert Tips for Application Readiness
NOTE: In terms of safety-critical deployments, ARM processors are becoming more and more prevalent in the aerospace, defense, and transportation industries. GNAT is a free-of-charge complete compiler for Ada95, integrated into the GCC compiler system.
The ARM platform adds to the GNAT Pro Safety-Critical product offering, which is already available for PowerPC and LEON boards, but now allowing portability among all three platforms.
NOTE: Bareboard refers to Printed Circuit Board technology that is supplied with conductors, but is not supplied with electronic components.
The technology here does not require any underlying operating system, so it can be deployed on very small memory boards. The tool suite includes support for Ada 2012 (including the "contract-based programming" features that make it easier to reflect the program's intent) and all earlier versions of the Ada language.
There is also support for the Ravenscar tasking profile, a set of static analysis tools, the GNATcheck coding standard verification tool, the GNATtest unit test harness generator, and the GDB visual debugger. A native Integrated Development Environment (IDE) as well as an Eclipse plug-in has also been provided.
AdaCore managing director Cyrille Comar says that GNAT Pro Safety-Critical for bareboard ARM supplies a fully configurable/customizable runtime library and implements high-integrity profiles that are especially relevant to safety-critical systems. "The Zero Footprint Profile (ZFP) in particular defines an Ada subset that does not require any runtime routines, thus reducing the memory footprint to user code only," he said.