Modern supercomputers can now reach a petaflop of processor power
Turnkey systems can be used in an innovative way to enable high performance computing (HPC) systems to work with high flexibility, scalability and performance, a new study as scientist at Virginia Tech University has claimed claimed.
In the modern world, high performance computers (HPC) systems or supercomputers play an important role in computer science. Researchers use these systems to solve complex issues that involve large amounts of data. for example in quantum mechanics, climate research and molecular modeling. The supercomputer can handle such issues in a very short time; Modern supercomputers can perform almost one quadrillion floating point operations per second (FLOPS), also known as a petaflop.
However, the storage platforms needed to build such HPC data systems are limited by a framework that requires users to choose between high availability or customization of features.
Researchers at Virginia Tech now claim that they have developed a new framework called BespoKV, which would allow researchers to build supercomputers that could perform billions of billion computations in one second (exascale computing).
BespoKV is based on the Key Value (KV) system, which stores important data on fast-memory storage instead of slower discs.
According to researchers, BespoKV's unique selling point is its ability to create several KV stores with desirable properties. It takes a "data" (single server KV store) and then creates immediate KV stores that are ready to use.
The researchers also argue that BespoKV eliminates the need to transform a system from the beginning to complete a specific task and say: "A developer can release a data in BespoKV and relieve the" messy pipe system "for distributed systems to the framework."
This study is relevant for industries that process large volumes of data, such as major credit card companies, movie-streaming sites, and social media.
"Developers from big companies can really lower their teeth to design innovative HPC storage systems with BespoKV," says Ali Butt, professor of computer science.
"Data access performance is a major limitation in HPC storage systems and usually uses a mix of solutions to provide flexibility along with performance, which is difficult. We have created a way to significantly speed up system behavior to meet desired performance, consistency, and reliability levels. "
The results of the study are presented today in the Association of Computing Machinery / IEEE Supercomputing Conference in Dallas, Texas.