I always get giddy about going to NVIDIA GPU Technology Conference (GTC), which is happening next week in San Jose. Just looking at the catalog of sessions gets me dreaming about what artificial intelligence will make possible and more importantly the business benefits that can already be derived from it.

I also love visiting all the startups on the show floor to see the latest innovations and the promises they offer of easier, better artificial intelligence frameworks to help garner faster, deeper insights.

But all this great software has to run on some hardware somewhere, be it in the cloud or on premises in your data center. Cisco offers a great array of infrastructure solutions to run your AI/ML projects. From the many iterations during the test/dev phase, to the learning and then the inferencing, we offer flexibility in your choice of platform. Come by our booth and see for yourself. We will have several Cisco UCS servers loaded with NVIDIA GPU on display and an array of demos. Here’s what we have planned:

AI Powered by Cisco UCS & Cisco HyperFlex

We will show how running Google Kubeflow on Cisco UCS or HyperFlex with NVIDIA GPU allows you to deploy your machine learning workloads on premises the same way you deploy them in the cloud. We will also demonstrate how you can provide additional GPU elasticity to your infrastructure for ML projects with Bitfusion running on Cisco UCS and Cisco HyperFlex.

Big Data with Deep Learning

Data preparation is usually a critical part of any deep learning project. This demo will show you how you can integrate your big data pipeline with a deep learning framework accelerated with the latest GPU.

AI Powered by NVIDIA Jetson

The Cisco Spark Board and Cisco Spark Room 55 are powerful video collaboration devices that bring more intelligence and usability to your small- and medium-sized rooms. Many of these features are delivered via artificial intelligence that Cisco has harnessed with the power of the NVIDIA Jetson chip, including:

– Real time video image processing of 4K images for intelligent camera views such as best overview and fast, active speaker tracking
– Face detection and people count and room usage analytics
– Fast image rendering on 4K display for real time white boarding – so that it’s as close to the analog experience as possible
– Automatic noise suppression that detects and reduces non-voice sounds

Win10 Full Throttle Graphics

Finally, another use for GPUs in the data center is for delivering virtualized graphics workstations. Our versatile Cisco UCS and HyperFlex servers equipped with the right GPUs are a great fit for running Citrix XenDesktop or VMware Horizon. We will get your mind racing and offer some fun with our heart-pounding virtual desktop demo where you can pit your skills against a colleague or a complete stranger.

Breakout session S8971: Adopting Artificial Intelligence in Networking

Come to Cisco’s session and get a chance to win an NVIDIA Shield TV. Hugo Latapie from our Chief Technology and Architecture Office will be discussing what is happening in the world of AI as it relates to networking, IT infrastructure, and IoT technologies. He will cover use cases and solutions ranging from smart city, transportation, manufacturing, to security and networking.

So make sure you stop by Cisco booth 415 and talk with our technical team about what the Cisco GPU accelerated data center can do for you.




Francoise Rees

Marketing Manager

Customer Solution Marketing, Cisco Intersight