Architecting for the Future: Intel Develops Innovative Mesh Interconnect Architecture

A pivotal objective of today’s computing systems architecture layout is to realize maximized performance across the wide range of usage situations. And, processors perform a crucial role in the effectiveness of computing systems, especially the choice of processor topology.

As such, Intel has been constantly reinventing its computing capabilities to enhance processor throughput and meet the growing demands for efficient systems.

Recently, the chip giant announced that it is designing an improved architecture for the forthcoming Xeon Scalable processors. The innovative architecture technology is tailored to reengineer the interconnection of on-chip parts, and magnify the competency and scalability of processors with several cores.

The Intel Xeon Scalable processors leverages a creative “mesh” on-chip interconnect topology that provides minimal latency, optimized bandwidth among various components, as well as enhanced scalability and performance.

Intel’s state-of-the-art mesh architecture organizes the various elements in rows and columns, with interior linking intersections to give room for turns.
Through designing a straighter pathway than the previous ring topology and an increased number of tracks to eradicate blockages, Intel’s new mesh can work with reduced frequency and voltage requirements while providing optimal bandwidth and minimal latency.

Consequently, it leads to enhanced performance and better energy usage—just like a properly constructed city road that allows traffic to move freely without any jamming.

Read the full article here.

Comments are closed.