NVIDIA Magnum IO is the architecture for parallel, intelligent data center I/O. It maximizes storage, network, and multi-node, multi-GPU communications for the world’s most important applications, using large language models, recommender systems, imaging, simulation, and scientific research. Magnum IO utilizes storage I/O, network I/O, in-network compute, and I/O management to simplify and speed up data movement, access, and management for multi-GPU, multi-node systems. It supports NVIDIA CUDA-X libraries and makes the best use of a range of NVIDIA GPU and networking hardware topologies to achieve optimal throughput and low latency. In multi-GPU, multi-node systems, slow CPU, single-thread performance is in the critical path of data access from local or remote storage devices. With storage I/O acceleration, the GPU bypasses the CPU and system memory, and accesses remote storage via 8x 200 Gb/s NICs, achieving up to 1.6 TB/s of raw storage bandwidth.