AMD unveils MI300x AI chip as 'generative AI accelerator'

To be able to run an entire 40-billion parameter model in memory, said Lisa Su, actually reduces the number of GPUs you need.

Related Keywords

San Francisco ,California ,United States ,Fairmont ,El Capitan ,Department Of Energy Lawrence Livermore National Laboratories ,Fairmont Hotel ,Open Compute Platform ,Grace Hopper ,Lawrence Livermore National Laboratories ,