A nanopass framework for commercial compiler development

Abstract

Contemporary compilers must typically handle sophisticated highlevel source languages, generate efficient code for multiple hardware architectures and operating systems, and support source-level debugging, profiling, and other program development tools. As a result, compilers tend to be among the most complex of software systems. Nanopass frameworks are designed to help manage this complexity. A nanopass compiler is comprised of many singletask passes with formally defined intermediate languages. The perceived downside of a nanopass compiler is that the extra passes will lead to substantially longer compilation times. To determine whether this is the case, we have created a plug replacement for the commercial Chez Scheme compiler, implemented using an updated nanopass framework, and we have compared the speed of the new compiler and the code it generates against the original compiler for a large set of benchmark programs. This paper describes the updated nanopass framework, the new compiler, and the results of our experiments. The compiler produces faster code than the original, averaging 15-27% depending on architecture and optimization level, due to a more sophisticated but slower register allocator and improvements to several optimizations. Compilation times average well within a factor of two of the original compiler, despite the slower register allocator and the replacement of five passes of the original 10 with over 50 nanopasses.

Pattern Eliminating Transformations

“…We also intend to extend and use the approach in the context of automatic rewrite rule generation techniques, such as the one introduced in [8], in order to automatize the generation of boilerplate code as in [22].…”

“…In particular, notions such as constructor subtypes [6] could be used to construct complex type systems whose type checking would provide guarantees similar to the ones provided by our formalism. This would however result in the construction of multiple type systems in order to type check each transformation as was the case in the original inspiration of our work [22].…”

“…Formalisms such as the one introduced for NanoPass [22] have proposed a method to eliminate a lot of the overhead induced by the denition of the intermediate languages by specifying only the symbols eliminated from the source language and generating automatically the corresponding intermediate language.…”

A type and scope safe universe of syntaxes with binding: their semantics and proofs

“…Moreover, it is possible to define generic transformations elaborating these added features in terms of lower-level ones. This suggests that this setup could be a good candidate to implement generic compilation passes and could deal with a framework using a wealth of slightly different intermediate languages à la Nanopass (Keep & Dybvig (2013)).…”

Schism: A Self-Hosting Scheme to WebAssembly Compiler

“…The first version to self-host can be seen on the Github repository under the Git tag self-host. 2 The compiler is structured as a multi-pass compiler inspired by the NanoPass framework [10]. Each pass takes a higher level form and translates it to a lower level version until finally we have a form that can be directly converted to WebAssembly.…”