Cisco has introduced its first Unified Compute System (UCS) server designed specifically to handle artificial intelligence (AI) and machine learning (ML) workloads. The Cisco UCS C480 ML is designed specifically for data scientists to perform AI and ML at every stage of the lifecycle.
It’s not like Cisco whipped up all kinds of special sauce for this server; it’s just a lot of very high-end components. The UCS C480 ML M5 rack server is a 4U device with the latest Intel Xeon processors and 8 Nvidia Tesla V100-32G GPUs with NVLink interconnects.
The top-of-the-line configuration features two Xeon processors, up to 128GB of DDR4 RAM, 24 SATA hard drives or SSDs, six NVMe SSD drives, and four x100G Virtual Interface Cards (VICs). The UCS C480 ML M5 is designed to work with Cisco’s various servers and HyperFlex systems with GPUs.
Cisco says it is covering each stage of the AI and ML lifecycle with this server. That ranges from data collection and analysis near the edge to data preparation and training in the data center, to the real-time inference at the heart of AI. It’s designed to let customers extract more intelligence from their data and use it to make better, faster decisions.
With its new DevNet AI Developer Center and DevNet Ecosystem Exchange, Cisco is also giving data scientists and developers the tools and resources to create a new generation of apps.
Cisco UCS features
There is actually quite a bit of software news around this announcement. The UCS server is designed to work with containerized apps and multicloud computing models for AI systems with data sets stored across multiple servers and services.
It’s fully compatible with Cisco’s AI solutions stack, which includes Cloudera’s Data Science Workbench, which supports frameworks such as TensorFlow and PyTorch, Hortonworks’ Hadoop 3.1, containerized Apache Spark and Google TensorFlow analytic workloads, and MapR data pipelines.