IBM PowerAI | AC922 Pre-configured server for Deep / Machine Learning

Introducing AC922 and IC922 systems. The world’s most powerful servers in 2U’s, designed by IBM specifically for deep learning and accelerated AI analysis in general. A system the result of the collaboration of IBM and NVIDIA will allow you to delv into the new level of AI and its applications: image recognition, natural language interpretation, medical diagnostics, social media analysis, etc.

Equipped with the latest NVIDIA graphics cards

V100 (up to 6 with liquid cooling or 4 by air). 2 TB of memory, NVLINK 2.0, PCIE 4 (world’s first supported server) and all major pre-compiled, optimized machine learning frameworks with official IBM support. They are the same machines that have been used in the most powerful super-computer in the world: SUMMIT Having gone from 18,688 X86 nodes to 4,600 POWER9 (1/4 machines) and still multiplied the power by between 5X and 10X depending on the application.

Google and Rackspace tthey have also joined the POWER9 Era through their Power9-based “ZAIUS” and “Barreleye G2” systems already operating successfully on their CPDs. If you have or plan to deploy Machine Learning solutions, this is your platform, request a demo right now.

Server features

  • 2 x IBM Power9 processor
  • 4 x In-memory bandwidth of Intel servers.
  • NVLink connectors, for direct connection of GPUs and sockets
  • 4 x Nvidia V100 cards designed specifically for Deep Learning
  • PowerAI*
  • Maximum density (2Us) and minimum power consumption
  • Optional enterprise support through IBM Deep Learning.

*PowerAI enables the deployment of infrastructures optimized for Artificial Intelligence solutions in hours. All optimized and adapted to the latest IBM hardware architectures.

Get the most out of machine learning technologies like IBM Watson Studio, RapidMiner, RStudio, and Qubole. All of them supported on IBM PowerAI on our AC922 and IC922 servers. The latter are cheaper and ideal for training or running small or non-critical production loads.

IBM AI servers give you a stable, robust, and tremendously durable platform that you’ll amortize in a few months compared to the costs of performing the same operations in the Azure, Google, or Amazon AWS clouds. Thousands of cybersecurity, image recognition, fraud detection, digital marketing, and production process automation projects already use these servers.

Work directly with our experts who will provide you with everything you need to get the most out of your applications. If you want to deploy a container-based hybrid cloud for ML, you may be interested in our OpenShift Container Platform-based ML solution.

© 2021 - SiXe Engineering | Training, consulting, professional services and turnkey projects | IBM, Lenovo, Docker, Red Hat , Tenable, HCL, MEDIGATE, Rapid7, Veeam, Sealpath & SUSE Authorized Business Partner. Company registered in INCIBE's cyber security catalog.
SiXe Ingeniería