We recently sat down with member company Microchip to discuss their relationship with the Gen-Z Consortium, their goals, and views on the industry. Read on for more.
Why did Microchip originally decide to join the Gen-Z Consortium?
Primarily because we are very engaged with infrastructure solutions for server and storage applications and the architectural transition that enables memory expansion for processors. We recognize that open standards, and industry collaboration, are key to the success for a new interface. That said, Gen-Z promises to be a robust, scalable option.
What are some of Microchip’s goals in working with Gen-Z technology and the Consortium?
Gen-Z is a memory fabric solution that enables heterogeneous memory and processing. With our expertise in high performance PCIe switching and advanced fabrics, our developments in data reliability in the SCSI domain, and our knowledge of the storage endpoint market, we are eager to contribute to the Consortium’s mission and to the definition of a Gen-Z memory architecture.
What do you feel is the importance of open-standard contributions ‒ such as Gen-Z ‒ for the HPC industry?
Microchip views open standards as critical for gaining momentum and broad adoption. Any point of interoperability requires clearly defined standards so every solution can successfully communicate with the fabric. Standards advance the technology by incorporating requirements from the many contributing members to deliver a robust, easily adopted and flexible interface. A clearly defined interface and strict interoperability compliance enable technology innovation.
What do you see as Gen-Z’s impact within the HPC industry?
Memory-centric computing is at the core of processing Big Data. Gen-Z expands and disaggregates byte addressable executable memory for every processing engine attached to the fabric. Separating memory from the processors has many benefits, including eliminating stranded or underutilized resources, reducing system-level costs by enabling dynamic provisioning of resources, enabling larger ‘execute in memory’ processes, improving performance and enabling applications to operate on larger scale databases, and enabling applications to scale and be executed across homogeneous processing platforms. This list is extensive and some benefits are immediately implementable while others will evolve as memory centric fabric architectures are adopted.
How do you see Gen-Z evolving to meet the needs of the industry?
Gen-Z currently resolves provisioning and resource allocation challenges for large data systems. As larger and faster memory is available to applications and processing workloads can be split across heterogeneous processing and accelerator engines, more performance gains will be enabled by the Gen-Z memory fabric and adjacent technologies such as CXL. Gen-Z enables rack scale systems that disaggregates resources across a fabric, helping to mitigate capital expense, power, cooling and replacement of components while allowing for dynamic scaling as need grow.
Be sure to follow our blog for more Gen-Z Consortium member spotlight features.