B blog.tensorwave.com AMD's MI300X Outperforms Nvidia's H100 for LLM Inference 280 points fvv 2 years ago 264 comments
