StarDBT: An efficient multi-platform dynamic binary translation system

This paper describes the design and implementation of a research dynamic binary translation system, StarDBT, which runs many real-world applications. StarDBT is a multi-platform translation system that is capable of translating application level binaries on either Windows or Linux OSes. A system-level variant of StarDBT can also run on a bare machine by translating the whole system code. We evaluate performance of a user-mode system using both SPEC2000 and some challenging Windows applications. StarDBT runs the SPEC2000 benchmark competitively to other state-of-the-art binary translators. For Windows applications that are typically multi-threaded GUI-based interactive applications with large code footprint, the StarDBT system provides acceptable performance in many cases. However, there are important scenarios in which dynamic translation still incurs significant runtime overhead, raising issues for further research. The major overheads are caused by the translation overhead of large volume of infrequently-executed code and by the emulation overhead for indirect branches.