Particle.news

Download on the App Store

GPUHammer Attack Prompts Nvidia to Mandate ECC on GDDR6 AI GPUs

Enabling ECC prevents silent AI model corruption at the cost of roughly 10 percent slower performance, a 6 percent reduction in usable VRAM

blank
Nvidia's Rowhammer problem has an easy fix.
Nvidia
Image

Overview

  • University of Toronto researchers publicly demonstrated GPUHammer, a Rowhammer-style exploit targeting GDDR6 memory.
  • Single-bit flips induced by GPUHammer can collapse AI accuracy from around 80 percent to under 1 percent on an RTX A6000.
  • GPUHammer requires only co-residency on the same GPU in cloud or server environments, avoiding any need to access user code or data.
  • Nvidia’s security advisory lists affected Ampere, Ada, Hopper and Turing models with a recommendation to activate ECC on GDDR6-based cards.
  • New Nvidia GPUs including the RTX 5090 and H100 integrate on-chip ECC to automatically guard against bit-flip attacks.