Combining Processor Virtualization and Split Compilation for Heterogeneous Multicore Embedded Systems

Complex embedded systems have always been heterogeneous multicore systems. Because of the tight constraints on power, performance and cost, this situation is not likely to change any time soon. As a result, the software environments required to program those systems have become very complex too. We propose to apply instruction set virtualization and just-in-time compilation techniques to program heterogeneous multicore embedded systems, with several additional requirements: * the environment must be able to compile legacy C/C++ code to a target independent intermediate representation; * the just-in-time (JIT) compiler must generate high performance code; * the technology must be able to program the whole system, not just the host processor. Advantages that derive from such an environment include, among others, much simpler software engineering, reduced maintenance costs, reduced legacy code problems... It also goes beyond mere binary compatibility by providing a better exploitation of the hardware platform. We also propose to combine processor virtualization with split compilation to improve the performance of the JIT compiler. Taking advantage of the two-step compilation process, we want to make it possible to run very aggressive optimizations online, even on a very constraint system.