Cloud Hosting

Political parties. Mind and body. Wealth and lack.

To the long list of divides society must navigate, add the growing gap between the capacity of hardware devices and the demands of the applications that run on them. But researchers like Huaicheng Li, assistant professor in the Department of Computer Science, and his team are working to solve this vexing computing challenge.

With applications that require more robust processing speed and efficiency, the pressure is on to make every server in every data center work harder. Under the auspices of a National Science Foundation Faculty Early Career Development Program (CAREER) award, Lis group will work to improve system efficiency, programmability, and performance to bridge the gap between emerging hardware technologies and the needs of modern software applications, Li said.

Over the past decade, increasing demands for improved data storage and processing capabilities have fueled rapid changes in the datacenter landscape.

Software such as the operating system plays a key role in hardware performance. Unfortunately, it has become clumsy. Everything reliant on high-level computing needs new technologies to enhance how memory, storage, and computing can be used. Lis group is working to expand the use with CXL specifically.

We envision a future where computer systems will adopt CXL for next generation storage, memory, and compute devices. Unfortunately, CXL is still very new, and it's not designed for nearly decade-old operating systems, Li said. Our ultimate goal is to introduce a holistic approach to unify memory, storage, and computing devices to make them faster, more cost-efficient, and easier to use.

This open-access project could enable advancements in data analytics, machine learning, and more. Its also expected to boost data center efficiency while reducing costs. Less expensive, high-performance services will benefit billions of end users, Li said.

Go here to read the rest:

Bridging the software-hardware divide - Virginia Tech