All news

InfrastructureAMD Handed Half of Inference to a Rival Chip

AMD Handed Half of Inference to a Rival Chip

On July 23 AMD and Cerebras built one inference pipeline from two rival chips: AMD Helios handles the prompt, the Cerebras wafer writes the answer. The 5x efficiency figure is a model, not a benchmark. Here is what disaggregated inference changes for your compute budget.

3 min read

Page 104 / 142