Researchers at the University of Toronto have developed a technique named GPUThor that can bypass error correction mechanisms on Nvidia GPUs. This vulnerability could potentially allow attackers to gain root shell access on affected hardware. AI
IMPACT This vulnerability could impact the security and reliability of AI workloads running on affected Nvidia GPUs.
RANK_REASON Research paper detailing a new technique impacting hardware. [lever_c_demoted from research: ic=1 ai=0.7]
Read on Mastodon — mastodon.social →
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →