A reflective functional language for hardware design and theorem proving

Abstract

This paper introduces reFL ect , a functional programming language with reflection features intended for applications in hardware design and verification. The reFL ect language is strongly typed and similar to ML, but has quotation and antiquotation constructs. These may be used to construct and decompose expressions in the reFL ect language itself. The paper motivates and presents the syntax and type system of this language, which brings together a new combination of pattern-matching and reflection features targeted specifically at our application domain. It also gives an operational semantics based on a new use of contexts as expression constructors, and it presents a scheme for compiling reFL ect programs into the λ-calculus using the same context mechanism.

Assume-guarantee validation for STE properties within an SVA environment

“…Forte is essentially a programming environment based around the reFL ect functional programming language [9], and large-scale STE verification efforts are to a great degree a programming (or scripting) activity in nature. STE model checking is invoked through reFL ect library function calls, and trajectory assertions for verification (lists of 5-tuples) are generated by writing functional programs that compute them.…”

“…The Forte implementation of STE is embedded in a functional programming language, reFL ect [9], similar to ML. STE antecedents are represented by lists of 5-tuples of the form (guard , node, value, start, end ) where guard and value are formulas of propositional logic (represented as BDDs), node is a node name (a string), and start and end are non-negative integers.…”

“…Our theoretical contribution is a simple solution to capturing the semantics of certain uses of STE symbolic indexing variables in the SVA language, where there is no similar concept, and in arguing the soundness of the translation. We also describe an implementation that employs the reflection features of the Forte system's functional scripting language [9] to allow both STE verification and translation from a single source. Finally, we present experimental results from extensive use of the framework on the STE verification environment of a recently taped-out Intel microprocessor.…”

A methodology for generating verified combinatorial circuits

“…Our work builds on a long tradition of using functional programming languages to describe hardware circuits, such as Ruby [12], Lava [2], Hawk [14], HML [18], Hydra [23], reFLect [9]. However, none of these languages provide the kind of manifest interfaces (static types) discussed above.…”

A multi-stage language with intensional analysis

“…This typing statically ensures that dynamically generated programs are type-safe, but excludes some functions that destruct or traverse the structure of expressions. Other approaches [20,8,22] assign the same type cod to all quoted expressions, performing their type checking at run-time. Such languages make a tradeoff between static and dynamic typing.…”