NVIDIA EGX Accelerates AI on TYAN Edge Server

Print Friendly, PDF & Email

Today TYAN announced that it is teaming with NVIDIA to integrate the NVIDIA EGX platform with its Thunder SX TN76-B7102 edge server, completing TYAN’s GPU server product line.

NVIDIA EGX provides the missing link for low-latency AI computing at the edge with an advanced, light-compute platform, reducing the amount of data that needs to be pushed to the cloud.

NVIDIA EGX is what we are looking for to solve the current bottleneck of cloud computing, which offers a robust computing power to AIoT and any other smart devices. We see strong demand from the enterprise market and are proud that TYAN Thunder SX TN76-B7102 is to be in the first wave of EGX partnerships” said Danny Hsu, Vice President of TYAN.

Featuring low-latency architecture between device and analytics endpoints, the TN76-B7102 EGX platform enables enterprises to efficiently process and respond to data. For example, a factory site can utilize the features of the EGX platform to react to the real-time response from any connected machines. This, in turn, reduces the amount of bandwidth needed compared to if that data had to be sent for processing all the way back to a data center or cloud.

TYAN’s platform combines a range of TYAN and NVIDIA AI technologies with security, networking, and storage architecture capabilities from Mellanox and other suppliers. With the TN76-B7102 server platform, companies in the largest industries—transportation, manufacturing, industrial, retail, healthcare, and agriculture—deploy AI quickly and securely, from edge to cloud.

TN76-B7102 is a 2U dual-socket 2nd gen Intel Xeon Scalable processor-based storage platform. With support for up to 4 NVIDIA T4, 24 DDR4 DIMM slots and maximum expansion up to 8 standard PCIe slots, the TN76-B7102 can fully address the high memory capacity required by data virtualization and in-memory databases and is ideal for Big Data.

The TN76-B7016 powered by NVIDIA EGX is now on display in booth #L0631a at Computex 2019 in Taiwan.

Sign up for our insideHPC Newsletter