GigaIO’s Power-efficient Interconnect Technology Achieves Breakthrough AI Performance: 2x Faster Training and Fine-Tuning with 83.5x Lower Latency

AI training, fine-tuning, and inference benchmark results include 2x faster training and fine-tuning and 83x better time to first token for inferencing with RDMA over RoCE.