The first results of my experiments with self-optimizing interpreters was finally published in IEEE Software. It is a brief and very high-level comparison of the Truffle approach with a classic bytecode-based interpreter on top of RPython. If you aren’t familiar with either of these approaches, the article is hopefully a good starting point. The experiments described in it use SOM, a simple Smalltalk.
Since writing things down, the work on the different SOM implementations has continued resulting in better overall performance. This reminds me: thanks again to the communities around PyPy/RPython and Truffle/Graal for their continues support!
With the rise of domain-specific languages (DSLs), research in language implementation techniques regains importance. While DSLs can help to manage the domain’s complexity, it is rarely affordable to build highly optimizing compilers or virtual machines, and thus, performance remains an issue. Ideally, one would implement a simple interpreter and still reach acceptable performance levels. RPython and Truffle are two approaches that promise to facilitate language implementation based on simple interpreters, while reaching performance of the same order of magnitude as highly optimizing virtual machines. In this case study, we compare the two approaches to identify commonalities, weaknesses, and areas for further research to improve their utility for language implementations.