A-QED Verification of Hardware Accelerators

Singh, Eshan; Lonsing, Florian; Chattopadhyay, Saranyu; Strange, Maxwell; Wei, Peng; Zhang, Xiaofan; Zhou, Yuan; Chen, Deming; Cong, Jason; Raina, Priyanka; Zhang, Zhiru; Barrett, Clark; Mitra, Subhasish

doi:10.1109/dac18072.2020.9218715

Cited by 10 publications

(10 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This paper presents both a formal foundation of A-QED 2 and an empirical evaluation that demonstrates its bug-finding capabilities in practice. We prove that A-QED's completeness guarantees [8] continue to hold for A-QED 2 -if the full HA under verification contains a bug, then A-QED 2 will detect that bug. Furthermore, we apply A-QED 2 to a wide variety of non-interfering LCAs (although our theoretical proofs apply to interfering LCAs as well): 109 different (buggy) versions of large open-source HAs of up to 200 million logic gates (including industrial HAs).…”

Section: Introductionmentioning

confidence: 88%

“…A-QED 2 builds on A-QED [8] and leverages BMC [11], [37]. Similar approaches based on self-consistency have been successfully applied to other classes of hardware designs, such as processor verification (as symbolic quick error detection (SQED) [38]- [43]), as well as to hardware security [44]- [49].…”

Section: Related Workmentioning

confidence: 99%

“…A recent formal verification technique targeting HAs, Accelerator-Quick Error Detection (A-QED) [8], overcomes the first challenge above. A-QED is readily applicable for a popular class of HAs: loosely-coupled accelerators (LCAs) [9], [10] (i.e., HAs that are not integrated as part of a central processing unit (CPU), but via an SoC's network-on-chip or a bus) that are also non-interfering.…”

Section: Introductionmentioning

confidence: 99%

“…Specifically, it checks for functional consistency (FC), the property that identical inputs within a sequence of operations always produce the same outputs. It was shown that FC checks, together with response bound (RB) checks and single-action correctness (SAC) checks, provide a thorough verification technique for non-interfering LCAs [8]. However, despite its success in discovering bugs in moderately-sized HA designs, A-QED suffers from the scalability challenges of formal tools.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Scaling Up Hardware Accelerator Verification using A-QED with Functional Decomposition

Chattopadhyay,

Lonsing,

Piccolboni

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

Hardware accelerators (HAs) are essential building blocks for fast and energy-efficient computing systems. Accelerator Quick Error Detection (A-QED) is a recent formal technique which uses Bounded Model Checking for pre-silicon verification of HAs. A-QED checks an HA for self-consistency, i.e., whether identical inputs within a sequence of operations always produce the same output. Under modest assumptions, A-QED is both sound and complete. However, as is well-known, large design sizes significantly limit the scalability of formal verification, including A-QED. We overcome this scalability challenge through a new decomposition technique for A-QED, called A-QED with Decomposition (A-QED 2 ). A-QED 2 systematically decomposes an HA into smaller, functional sub-modules, called sub-accelerators, which are then verified independently using A-QED. We prove completeness of A-QED 2 ; in particular, if the full HA under verification contains a bug, then A-QED 2 ensures detection of that bug during A-QED verification of the corresponding subaccelerators. Results on over 100 (buggy) versions of a wide variety of HAs with millions of logic gates demonstrate the effectiveness and practicality of A-QED 2 . This article will appear in the proceedings of Formal Methods in Computer-Aided Design (FMCAD 2021).

show abstract

Section: Introductionmentioning

confidence: 88%

Section: Related Workmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Scaling Up Hardware Accelerator Verification using A-QED with Functional Decomposition

Chattopadhyay,

Lonsing,

Piccolboni

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…SQED employs the principle of self-consistency based on a mathematical interpretation of instructions as functions. That principle is also applied by accelerator quick error detection (A-QED) [24], a formal pre-silicon verification technique for HW accelerator designs. A-QED checks the functions implemented by an accelerator for functional consistency and, like SQED, does not require a formal specification.…”

Section: Related Workmentioning

confidence: 99%

A Theoretical Framework for Symbolic Quick Error Detection

Lonsing,

Mitra,

Barrett

2020

Preprint

Self Cite

View full text Add to dashboard Cite

Symbolic quick error detection (SQED) is a formal pre-silicon verification technique targeted at processor designs. It leverages bounded model checking (BMC) to check a design for counterexamples to a self-consistency property: given the instruction set architecture (ISA) of the design, executing an instruction sequence twice on the same inputs must always produce the same outputs. Self-consistency is a universal, designindependent property. Consequently, in contrast to traditional verification approaches that use design-specific assertions (often generated manually), SQED does not require a formal design specification or manually-written properties. Case studies have shown that SQED is effective on commercial designs and substantially improves design productivity. However, until now there has been no formal characterization of its bug-finding capabilities. We aim to close this gap by laying a formal foundation for SQED. We use a transition-system processor model and define the notion of a bug using an abstract specification relation. We prove the soundness of SQED, namely that any bug reported by SQED is in fact a real bug in the processor. Importantly, this result holds regardless of what the actual specification relation is. We next describe conditions under which SQED is complete, that is, what kinds of bugs it is guaranteed to find. We show that for a large class of bugs, SQED can always find a trace exhibiting the bug. Our results enable a rigorous understanding of SQED and its bug-finding capabilities and give insights on how to optimize implementations of SQED in practice.

show abstract