Heiko Becker

Elementary function operations such as sin and exp cannot in general be computed exactly on today’s digital computers, and thus have to be approximated. The standard approximations in library functions typically provide only a limited set of precisions, and are too inefficient for many applications. Polynomial approximations that are customized to a limited input domain and output accuracy can provide superior performance. In fact, the Remez algorithm computes the best possible approximation for a given polynomial degree, but has so far not been formally verified. This paper presents Dandelion, an automated certificate checker for polynomial approximations of elementary functions computed with Remez-like algorithms that is fully verified in the HOL4 theorem prover. Dandelion checks whether the difference between a polynomial approximation and its target reference elementary function remains below a given error bound for all inputs in a given constraint. By extracting a verified binary with the CakeML compiler, Dandelion can validate certificates within a reasonable time, fully automating previous manually verified approximations.

Verified compilers such as CompCert and CakeML have become increasingly realistic over the last few years, but their support for floating-point arithmetic has thus far been limited. In particular, they lack the “fast-math-style” optimizations that unverified mainstream compilers perform. Supporting such optimizations in the setting of verified compilers is challenging because these optimizations, for the most part, do not preserve the IEEE-754 floating-point semantics. However, IEEE-754 floating-point numbers are finite approximations of the real numbers, and we argue that any compiler correctness result for fast-math optimizations should appeal to a real-valued semantics rather than the rigid IEEE-754 floating-point numbers. This paper presents RealCake, an extension of CakeML that achieves end-to-end correctness results for fast-math-style optimized compilation of floating-point arithmetic. This result is achieved by giving CakeML a flexible floating-point semantics and integrating an external proof-producing accuracy analysis. RealCake’s end-to-end theorems relate the I/O behavior of the original source program under real-number semantics to the observable I/O behavior of the compiler generated and fast-math-optimized machine code. The artifact is available at https://doi.org/10.4230/DARTS.8.2.10

Proof engineering efforts using interactive theorem proving have yielded several impressive projects in software systems and mathematics. A key obstacle to such efforts is the requirement that the domain expert is also an expert in the low-level details in constructing the proof in a theorem prover. In particular, the user needs to select a sequence of tactics that lead to a successful proof, a task that in general requires knowledge of the exact names of a large set of tactics.

We present Lassie, a tactic framework for the HOL4 theorem prover that allows individual users to define their own tactic language by example, and for instance give frequently used tactics or tactic combinations easier-to-remember names. The core of Lassie is an extensible semantic parser, which allows the user to interactively extend the tactic language through a process of definitional generalization. Defining tactics in Lassie thus does not require any knowledge in implementing custom tactics, while proofs written in Lassie retain the correctness guarantees provided by the HOL4 system. We show through case studies how Lassie can be used in small and larger proofs by novice and more experienced interactive theorem prover users, and how we envision it to ease the learning curve in a HOL4 tutorial.

When compared to idealized, real-valued arithmetic, finite precision arithmetic introduces unavoidable errors, for which numerous tools compute sound upper bounds. To ensure soundness, providing formal guarantees on these complex tools is highly valuable. In this paper we extend one such formally verified tool, FloVer. First, we extend FloVer with an SMT-based domain using results from an external SMT solver as an oracle. Second, we implement interval subdivision on top of the existing analyses. Our evaluation shows that these extensions allow FloVer to efficiently certify more precise bounds for nonlinear expressions.

Verified compilers like CompCert and CakeML offer increasingly sophisticated optimizations. However, their deterministic source semantics and strict IEEE 754 compliance prevent the verification of “fast-math” style floating-point optimizations. Developers often selectively use these optimizations in mainstream compilers like GCC and LLVM to improve the performance of computations over noisy inputs or for heuristics by allowing the compiler to perform intuitive but IEEE 754-unsound rewrites. We designed, formalized, implemented, and verified a compiler for Icing, a new language which supports selectively applying fast-math style optimizations in a verified compiler. Icing’s semantics provides the first formalization of fast-math in a verified compiler. We show how the Icing compiler can be connected to the existing verified CakeML compiler and verify the end-to-end translation by a sequence of refinement proofs from Icing to the translated CakeML. We evaluated Icing by incorporating several of GCC’s fast-math rewrites. While Icing targets CakeML’s source language, the techniques we developed are general and could also be incorporated in lower-level intermediate representations.

Being able to soundly estimate roundoff errors in finite-precision computations is important for many applications in embedded systems and scientific computing. Due to the unintuitive nature of finite-precision arithmetic, automated static analysis tools are highly valuable for this task. The results, however, are only as correct as the implementations of the static analysis tools. This paper presents a formally verified and modular tool which fully automatically checks the correctness of finite-precision roundoff error bounds encoded in a certificate. We present implementations of certificate generation and checking for both Coq and HOL4 and evaluate it on a number of examples from the literature. The experiments use both in-logic evaluation of Coq and HOL4, and execution of extracted code outside of the logics. We benchmark Coq extracted unverified OCaml code and a CakeML-generated verified binary.

Recent renewed interest in optimizing and analyzing floating-point programs has lead to a diverse array of new tools for numerical programs. These tools are often complementary, each focusing on a distinct aspect of numerical programming. Building reliable floating point applications typically requires addressing several of these aspects, which makes easy composition essential. This paper describes the composition of two recent floating-point tools; Herbie, which performs accuracy optimization, and Daisy, which performs accuracy verification. We find that the combination provides numerous benefits to users, such as being able to use Daisy to check whether Herbie’s unsound optimizations improved the worst-case roundoff error, as well as benefits to tool authors, including uncovering a number of bugs in both tools. The combination also allowed us to compare the different program rewriting techniques implemented by these tools for the first time. The paper lays out a road map for combining other floating-point tools and for surmounting common challenges.