In my first blog this year, I discussed Cisco UCS X-Fabric Technology and adding GPUs with the new PCIe nodes. What if you don’t need four GPUs but one or two Nvidia T4 GPUs would be perfect for your solution? UCS X-Series, just like our C-Series Rack Servers, support GPUs on the compute node itself.
It wasn’t just the UCS X9508 Chassis that was designed with modularity in mind, so was the UCS X210c Compute Node. When the UCS X210c initially launched, we had optional storage front mezzanine adapters. These modular, flexible options allow you the freedom to choose the right performance / cost solution. If you are booting from SAN and don’t need local storage, there is a solution. If you need storage, there is a high-performance pass through all NVMe option and a more traditional SAS/SATA option with a LSI RAID controller. Both support up to six drives.
What if your application and user experience could be enhanced by adding GPUs and you need to maintain the maximum CPU/memory density within your X-Series deployment? You now have the flexibility to add GPUs with a new mezzanine option on the compute node itself.
This front mezzanine adapter supports two NVMe drives and one or two Nvidia T4 GPUs. The T4 GPU excels at a wide range of modern applications, including machine learning, deep learning, data analytics, video transcoding, and virtual desktops (VDI). GPUs can help you optimize the performance of your applications running on the UCS X210c.
Like with the UCS X-Fabric Technology based GPUs, the on-server GPUs are fully manageable from Intersight. You can see which nodes have GPUs and have specific profiles for those servers. If a firmware update is needed, you won’t have to use a separate utility. Like all UCS firmware components, you’ll be able to update it within Intersight. This means a single tool and process can be put in place regardless of the server use case or components.
A single operational model for all your workloads
Cisco believes you shouldn’t have to create hardware silos to meet the needs of an application and provide your users the best experience. When X-Series launched last June, it supported a wide range of applications including some that need a rack server form factor because of local storage requirements. Now with Cisco UCS X-Fabric Technology and GPU announcements, we are expanding the portfolio to address a much wider range of capabilities to tackle the most demanding of applications while offering the modularity needed to face future demands without compromise.
When combined with UCS C-Series, Cisco HyperFlex, our converged infrastructure partners, and Cisco Intersight, you have a single operational model regardless of the form factor that works best. No one besides Cisco offers this.
Cisco UCS X210c M6 Compute Node
Explore all the UCS X-Series blogs