CESGA Deploys Quside QRNG Into Their Datacenter
Santiago de Compostela — The availability of large quantities of high-quality, high-performance random numbers has become an essential requirement in many fields. Its application ranges from high-quality encryption keys, which encrypt the exponential growth of data transfers, to the creation of highly accurate and time-efficient simulations for finance, insurance, engineering, and science.
QRNG to overcome limitations
Many computational problems and workloads from both industry and research rely on stochastic processes, which require an ever-increasing amount of random numbers. Limitations in current means for generating these random numbers lead to reduced performance, inefficient use of compute resources and even potential artifacts in some instances, especially in highly-parallel simulations and simulations of very sensitive quantities. Unfortunately, current data center and supercomputing infrastructures do not have access to high-quality and high-performance random number generators, and their source of physical randomness is either inexistent or very slow.
To overcome the aforementioned difficulties and inefficiencies when running stochastic workloads, CESGA has recently deployed a QRNG from Quside, to offer a unique and more powerful source of random numbers to their users.
Performance improvements and new capabilities
By leveraging Quside’s QRNG, randomized workloads such as Monte Carlo simulations or a variety of physics-inspired simulations can obtain performance improvements of up to 10x, as well as more accurate results thanks to the acceleration on the provision and processing of randomness.
Quside has proven up to multi-Gbps performance rates using its proprietary phase-difusion quantum random number generation technology and a range of hardware acceleration tools. In addition Quside provides proof of origin entropy metrics through their Randomness Metrology module. With this installation, CESGA users can start leveraging this unique new capability in their stochastic workloads.