In his virtual presentation at the start of the MWC Barcelona 2022, Intel has announced its advances in advanced computing, and has shown new programmable hardware and open software. These two advancements will drive untold innovation at scale and enable Intel to grow and broaden its leadership.
Have been submitted architectural enhancements and future central processing units on its next generation Intel Xeon Scalable processor, called Sapphire Rapids; and a new Intel Xeon D processor. The latter is Intel's latest SoC designed from the ground up for networking and edge software defined. They also announced enhancements to their software such as OpenVINO 2022.1, new software modules in Intel Smart Edge, to increase performance and flexibility at the edge and for developers.
Sapphire Rapids to improve vRAN workloads
Intel is redoubling its commitment to continue the growth of vRAN ecosystems to deliver uncompromising performance that meets key operator metrics. They are working with Samsung, Ericsson, Rakuten Symphony and other vendors to bring new solutions to market based on their Xeon Scalable platform.
Some information that they have made public is the new and exclusive improvements in the specific signal processing instructions for 5G in the Sapphire Rapids core. They have been created to support RAN-specific signal processing, they will offer up to a two-fold increase in capacity for vRAN and will support advanced capabilities such as high cell density for 64T64R Massive MIMO. This will allow customers to deploy vRAN in the world's most demanding environments.
The Sapphire Rapids family of server processors will be completed with new integrated acceleration chips optimized for vRAN workloads. They will grant greater flexibility and more optimization points when building your networks. There will be a wide selection of options in the next generation Intel Xeon Scalable processor family.
New Intel SoC Designed for Software Defined Networking and the Edge
Intel's new Xeon D processor has been designed from the ground up and is packed with networking-specific capabilities, not forgetting edge computing. This new Intel Xeon D It offers built-in AI and crypto acceleration, built-in Ethernet, support for time-coordinated computing and time-aware networking, and industrial-class reliability.
It enables low-latency, high-performance computing, with more security and network enhancements at the edge when installed on certain. operating environments. Some of these environments include sites that have space and power limitations. 70 technology companies are working with Intel on designs using the latest Intel Xeon D processors, including Cisco, Juniper Networks and Rakuten Symphony.
Edge computing and the rise of AI inference have become a critical workload for many industries. To this end, Intel has announced two software updates, with a strong edge developer experience. They want to make sure that Intel's hardware capabilities are accessible to developers in a cohesive way and that it's easy to program across modern, open cloud-native platforms.
To help developers continue to advance AI and take advantage of the performance of new and existing hardware, they have released a new OpenVINO update with three more to come. The first is a Updated and simplified API to more easily import TensorFlow models for improved code portability and faster time to market.
An expansion of model coverage follows, with increased support for natural language processing, double-precision models, and advanced computer vision. With this they create a significant performance boost across a wider range of use cases and model types. Finally, there is a new auto-optimization capability that automatically discovers all accelerators and computers on a given system and then dynamically balances the load and increases AI parallelization based on memory and compute capacity. Thus, application developers no longer need to know in advance what the hardware configuration will be because the software will adapt to them.
They have also announced new software modules in their Smart Edge portfolio. They are optimized on Intel Xeon processors to accelerate 5G User Plane Function workloads at the edge of the network. Thus they help to offer the required latency and bandwidth. These modules abstract hardware complexities so developers can write applications on top that take advantage of the packet processing capabilities of Intel CPUs, providing an easier path to improve 5G UPF performance.