arXiv Science⌕ Search
Home/Archive/arXiv

arXiv · 2603.12485

Hunting CUDA Bugs at Scale with cuFuzz

Mohamed Tarek Ibn ziad·Christos Kozyrakis
View original source ↗Read source document ↗Export metadata ↓

Abstract

GPUs play an increasingly important role in modern software. However, the heterogeneous host-device execution model and expanding software stacks make GPU programs prone to memory-safety and concurrency bugs that evade static analysis. While fuzz-testing, combined with dynamic error checking tools, offers a plausible solution, it remains underutilized for GPUs. In this work, we identify three main obstacles limiting prior GPU fuzzing efforts: (1) kernel-level fuzzing leading to false positives, (2) lack of device-side coverage-guided feedback, and (3) incompatibility between coverage and sanitization tools. We present cuFuzz, the first CUDA-oriented fuzzer that makes GPU fuzzing practical by addressing these obstacles. cuFuzz uses whole program fuzzing to avoid false positives from independently fuzzing device-side kernels. It leverages NVBit to instrument device-side instructions and merges the resultant coverage with compiler-based host coverage. Finally, cuFuzz decouples sanitization from coverage collection by executing host- and device-side sanitizers in separate processes. cuFuzz uncovers 43 previously unknown bugs (19 in commercial libraries) across 14 CUDA programs, including illegal memory accesses, uninitialized reads, and data races. cuFuzz achieves significantly more discovered edges and unique inputs compared to baseline approaches, especially on closed-source targets. Moreover, we quantify the execution time overheads of the different cuFuzz components and add persistent-mode support to improve the overall fuzzing throughput. Our results demonstrate that cuFuzz is an effective and deployable addition to the GPU testing toolbox. cuFuzz is publicly available at https://github.com/NVlabs/cuFuzz/.

Explore related subjects

cs.CRcs.SE

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Mohamed Tarek Ibn ziad, Christos Kozyrakis. 2026-03-12. Hunting CUDA Bugs at Scale with cuFuzz. https://doi.org/10.1145/3798231

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related papers

arXiv2026

SynGhost: Invisible and Universal Task-agnostic Backdoor Attack via Syntactic Transfer

Although pre-training achieves remarkable performance, it suffers from task-agnostic backdoor attacks due to vulnerabilities in data and training mechanisms. These attacks can transfer backdoors to various downstream tasks. In this paper, we introduce $\mathtt{maxEntropy}$, an entropy-based poisoning filter that mitigates such risks. To overcome the limitations of manual target setting and explicit triggers, we propose $\mathtt{SynGhost}$, an invisible and universal task-agnostic backdoor attack via syntactic transfer, further exposing vulnerabilities in pre-trained language models (PLMs). Specifically, $\mathtt{SynGhost}$ injects multiple syntactic backdoors into the pre-training space through corpus poisoning, while preserving the PLM's pre-training capabilities. Second, $\mathtt{SynGhost}$ adaptively selects optimal targets based on contrastive learning, creating a uniform distribution in the pre-training space. To identify syntactic differences, we also introduce an awareness module to minimize interference between backdoors. Experiments show that $\mathtt{SynGhost}$ poses significant threats and can transfer to various downstream tasks. Furthermore, $\mathtt{SynGhost}$ resists defenses based on perplexity, fine-pruning, and $\mathtt{maxEntropy}$. The code is available at https://github.com/Zhou-CyberSecurity-AI/SynGhost.

cs.CR↗
arXiv2026

Towards the ideals of Self-Recovery and Metadata Privacy in Social Vault Recovery with Apollo

Social recovery enables users to enlist trusted contacts, or trustees, to help recover lost access to end-to-end-encrypted repositories or vaults. However, existing recovery mechanisms often make strong memorability assumptions about what users will remember. Weakening these memorability assumptions to increase the robustness of recovery is possible, but may leak sensitive metadata about the user and/or trustees if done naively. This paper's first contribution is to draw attention to and formalize this basic tension between memorability and metadata privacy in social vault recovery. Our second contribution is Apollo, a social recovery mechanism that aims to avoid any memorability assumptions while strongly protecting recovery metadata privacy. Apollo approximates the ideal of self-recovery by relying only on a threshold of social reconnection events, which may be initiated either by the user or the user's contacts. Apollo thereby has a chance of succeeding in vault recovery even in a worst-case scenario where the user has forgotten all metadata, including even the vault's existence. To protect the metadata's privacy, Apollo distributes either real or fake (chaff) data to all of a user's contacts, not just the user's trustees, thus systematically anonymizing the trustees among the larger set of contacts. To make this anonymity set scalable, Apollo uses a novel multi-layered secret sharing scheme to mitigate the computational overhead of recovery in this setting, which would otherwise be exponential in the recovery threshold. Finally, we evaluate a prototype implementation of Apollo. Apollo reduces the probability of malicious recovery to under 0.1% for an adversary capable of obtaining shares from every 1-out-of-2 contacts. After reconnecting with 30 contacts, the multi-layered design shows an improvement of 5 orders in computation time, compared to a single-layered approach.

cs.CR↗
arXiv2026

Computational Certified Deletion Property of Magic Square Game and its Application to Classical Secure Key Leasing

We present the first construction of a computational Certified Deletion Property (CDP) achievable with classical communication, derived from the compilation of the non-local Magic Square Game (MSG). We leverage the KLVY compiler to transform the non-local MSG into a 2-round interactive protocol, rigorously demonstrating that this compilation preserves the game-specific CDP. Previously, the quantum value and rigidity of the compiled game were investigated. We emphasize that we are the first to investigate CDP (local randomness in [Fu and Miller, Phys. Rev. A 97, 032324 (2018)]) for the compiled game. Then, we combine this CDP with the framework [Kitagawa, Morimae, and Yamakawa, Eurocrypt 2025] to construct Secure Key Leasing with classical Lessor (cSKL). SKL enables the Lessor to lease the secret key to the Lessee and verify that a quantum Lessee has indeed deleted the key. In this paper, we realize cSKL for PKE, PRF, and digital signature. Compared to prior works for cSKL, we realize cSKL for PRF and digital signature for the first time. In addition, we succeed in weakening the assumption needed to construct cSKL.

cs.CR↗