Introducing AIRI from Pure Storage
Architected by Pure Storage and NVIDIA, enabling AI-at-scale for every enterprise
Introducing the AI-ready Infrastructure. AIRI.
Enterprises will tap the power of AI for faster innovation and a competitive edge. Yet AI requires a completely new infrastructure, and the complexities of legacy solutions are holding enterprises back from moving into the new era of intelligence. Scan is proud to partner with Pure Storage to offer the complete AIRI solutions to deliver best-in-class Ai and deep learning performance. AI requires new, modern technologies like GPUs, scale-out flash, and RDMA fabric to move tremendous amounts of data. Unfortunately, too many AI initiatives are stalled with the complexities of a “do-it-yourself” approach using legacy technologies, leading to months of delays and idle time.
AIRI
4x NVIDIA DGX-1 SYSTEMS
4 PFLOPS of DL Performance
PURE FLASHBLADE
15x 17TB Blades 1.5M NFS IOPS
CONVERGED FABRIC
2x 100GB Ethernet Switches with RDMA
Software
NVIDIA GPU CLOUD DEEP LEARNING STACK
NVIDIA Optimised Frameworks
AIRI Mini
2x NVIDIA DGX-1 SYSTEMS
2 PFLOPS of DL Performance
PURE FLASHBLADE
7x 17TB Blades 700K NFS IOPS
CONVERGED FABRIC
2x 100GB Ethernet Switches with RDMA
Software
AIRI SCALING TOOLKIT
Multi-Node Training Made Simple
The Core of AIRI
AI pushes beyond the reach of legacy technologies like serial CPUs and spinning disks. At the core of AIRI are NVIDIA® DGX-1™ servers and FlashBlade, industry-leading solutions architected for AI. Each replaces the performance of many racks of legacy technology, offering data scientists the power of a large supercomputer for any AI initiative – in a converged solution with a fraction of the physical space and power and cooling costs of legacy infrastructure.
The AIRI Technology Stack
FAST
AIRI powers multi-node deep learning out-of-the-box, delivering linear-scale performance for critical training workloads. It’s built on state-of-the-art technologies like RDMA over 100Gb Ethernet and leading software tools.
SIMPLE
Get started in hours, not weeks. NVIDIA GPU Cloud Deep Learning Stack delivers optimized frameworks immediately – while the AIRI Scale-out Training Kit enables a quick start on multi-DGX-1 training workloads.
FUTURE-PROOF
The most advanced platform for real-world AI, AIRI also offers the simplest scalable architecture for demands at any scale. Add GPUs for faster training, and/or add blades for bigger datasets. AIRI grows with your AI workload needs – without downtime.
AIRI is built with a complete software stack enabling data scientists to get up and running in a few hours, not weeks or months. AIRI Scaling Toolkit allows users to run their first multi-DGX training within hours, with a few commands and to slash training time from weeks to days for their most critical AI workloads.
AI-Ready Infrastructure Technical Data
Scan Computers, partnering with Pure Storage and NVIDIA are able to provide proof of concept testing, deployment planning, configuration advice and help with scaling these innovative AI architectures. Please contact our team to understand further how the AIRI range of solutions can revolutionise your AI and deep learning projects.
View Technical Document