NVIDIA introduced NVIDIA Magnum IO, a suite of software to help data scientists and AI and high performance computing researchers process massive amounts of data in minutes, rather than hours. This offering ahs been developed in close collaboration with vendors in networking and storage, including DataDirect Networks, Excelero, IBM, Mellanox and WekaIO.
Optimized to eliminate storage and input/output bottlenecks, Magnum IO delivers up to 20 times faster data processing for multi-server, multi-GPU computing nodes when working with massive datasets to carry out complex financial analysis, climate modeling and other HPC workloads.
At the heart of Magnum IO is GPUDirect, which provides a path for data to bypass CPUs and travel on “open highways” offered by GPUs, storage and networking devices. Compatible with a wide range of communications interconnects and APIs — including NVIDIA NVLink and NCCL, as well as OpenMPI and UCX — GPUDirect is composed of peer-to-peer and RDMA elements.
Its newest element is GPUDirect Storage, which enables researchers to bypass CPUs when accessing storage and quickly access data files for simulation, analysis or visualization.
Magnum IO integrates infrastructure elements to maximize storage and network I/O performance and functionality. It optimizes IO performance by bypassing the CPU to enable direct IO between GPU memory and network storage. It also relieves CPU contention to create a more balanced GPU-accelerated system and delivers peak IO bandwidth with up to ten times fewer CPU cores, and provides optimized implementation for current and future platforms, whether the data transfers are latency sensitive, bandwidth sensitive, or collectives.
NVIDIA Magnum IO software is available, with the exception of GPUDirect Storage, which is currently available to select early-access customers. Broader release of GPUDirect Storage is planned for the first half of next year.
No comments:
Post a Comment