Powered by OpenAIRE graph
Found an issue? Give us feedback
image/svg+xml art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos Open Access logo, converted into svg, designed by PLoS. This version with transparent background. http://commons.wikimedia.org/wiki/File:Open_Access_logo_PLoS_white.svg art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos http://www.plos.org/ ZENODOarrow_drop_down
image/svg+xml art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos Open Access logo, converted into svg, designed by PLoS. This version with transparent background. http://commons.wikimedia.org/wiki/File:Open_Access_logo_PLoS_white.svg art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos http://www.plos.org/
ZENODO
Conference object
Data sources: ZENODO
addClaim

SABRE: Deterministic Post-Silicon Microarchitectural Audit via Symmetric Control Flow Isolation

Authors: Aouini, Radouane;

SABRE: Deterministic Post-Silicon Microarchitectural Audit via Symmetric Control Flow Isolation

Abstract

This paper presents SABRE (Symmetric Architecture & Boundary Risk Evaluator), a deterministic bare-metal diagnostic framework designed for high-precision microarchitectural metrology on modern out-of-order processors. While traditional post-speculative hardware security evaluations rely on probabilistic fuzzing that suffers from severe environmental noise—such as arithmetic calculation bias, spatial prefetcher triggers, and frontend instruction cache decoding noise—SABRE introduces Symmetric Control Flow Isolation (SCFI). By forcing the CPU to execute mathematically identical machine instruction sequences in both low and high-entropy evaluation campaigns, SABRE isolates pure speculative execution delays with nanosecond-level precision. Evaluated on Intel Xeon (Sandy Bridge-EP) and AMD EPYC (Zen 3) microarchitectures, the framework successfully detects Reorder Buffer (ROB) saturation cliffs and quantifies genuine cross-thread ROB partitioning under SMT/Hyper-Threading contention directly on host silicon without relying on statistical noise-filtering classifiers.

Powered by OpenAIRE graph
Found an issue? Give us feedback