We propose to search for common abstractions for concurrency models to enable multi-language virtual machines to support a wide range of them. This would enable domain-specific solutions for concurrency problems. Furthermore, such an abstraction could improve portability of virtual machines to the vastly different upcoming many-core architectures. Problem Statement Since the processor manufacturers reached the boundaries of what is feasible to achieve computational speedups in terms of

more »

... clock rates, they changed their scaling dimension from clock rate to core count, i. e., the number of computing units on a single chip. With this change, they are still able to deliver more computing power with every new processor generation by shifting the burden of realizing speedups to the software developers [12] . In the fields of operating systems, middleware, databases, and distributed systems, solutions have been found to provide support for concurrency. However, for the development of end-user applications, today's systems still lack comprehensive support to make concurrency accessible and its complexity manageable. The most widely used programming model for concurrency is the shared memory model with threads and locks. Unfortunately, this model has very narrow limits. Even though threading and finegrained locking have proven to allow high-performance concurrency, programming complexity increases fast with a rising number of threads and shared resources. Therefore, this model does not scale up to the degree of concurrency in many-core system. The use of disciplined concurrency models avoiding shared state, is almost the only choice to cope with the inherent complexity. With the actor model and software transactional memory, there are promising candidates available, but currently not part of mainstream software de-