Technical Report Number
Computer Systems Organization, Computing Methodologies, Software
Traditional dynamic memory management techniques for imperative programming languages are unsuitable for reliable real-time applications because their worst-case time and space requirements are insufficiently bounded. This is demonstrated by detailed measurements of several real-world workloads. A special hardware-assisted real-time garbage collection system has been designed to facilitate reliable use of dynamic memory in hard real-time systems. By analyzing the dynamic memory use of application software, the real-time developer can prove compliance with time and space constraints. Analysis techniques are presented and the real-time performance of the hardware-assisted garbage collection system is compared to that of traditional allocators.