GPUHammer — a Rowhammer attack variant on GPU memory
Researchers from the University of Toronto have presented the first Rowhammer-class attack used to distort graphics memory content. The ability to execute an attack that results in the substitution of up to 8 bits of data has been demonstrated on the discrete NVIDIA A6000 GPU with GDDR6 memory. A practical example shows how such distortion can be used to interfere with the execution of machine learning models and substantially (from 80% to […]
