SANTA CLARA, Calif.--(EON: Enhanced Online News)--CORNAMI, a high-performance computing company in artificial intelligence (AI), machine learning, and big data, today announced that key monolithic patents around its next generation, highly-efficient, advanced, multi-core architecture technology have been issued, thereby greatly enhancing the CORNAMI IP asset value portfolio. The CORNAMI patent portfolio now has over 60 patents with more than another dozen pending in US and International PTO (Patent and Trademark offices).
“Building on our early breakthrough in multi-core processing technologies has enhanced CORNAMI’s position in developing the key ingredients to develop an advanced next-generation parallel architecture”
CORNAMI has developed a non-Von Neumann parallel architecture with independent decision making capabilities at each processing core, interspersed with high-speed memory, all interconnected by a biologically inspired network to produce a scalable “sea of cores”. This unique architecture delivers tremendous advancements in efficient multi-core parallel processing that dramatically changes the output-to-power performance at the petabyte data-set scale. There is built-in demand for real-time and actionable data analytics for applications in the hyper-growth big data, machine learning and AI markets. This technology significantly increases performance while greatly reducing power, latency, footprint and costs, resulting in greater overall compute efficiencies.
Other key advantages of this unique architecture include faster-time-to-market; increase in profits from energy savings and labor efficiencies; and scalability – providing an almost infinitely extensible fabric of cores without the traditional overhead on current architectures, systems and computing clusters.
“The world’s immense processing workloads have dramatically risen within the last five years, driving the need for a next-generation architecture,” stated Paul Master, co-founder and CTO at CORNAMI. “Compute-intensive markets – such as machine learning, AI, big data -- need a way to parallelize processing-workloads to run across a plethora of small cores, (effectively a 'sea of cores') to achieve greater efficiency and deliver real-time and predictive analysis for these high-value markets.”
“Building on our early breakthrough in multi-core processing technologies has enhanced CORNAMI’s position in developing the key ingredients to develop an advanced next-generation parallel architecture,” said Gordon Campbell, co-founder and CEO of CORNAMI. “The industry is in need of closing the gap between the exploding growth in big data and the exponential yearly growth of transistors per chip with the leveling out of CPU performance over the last decade. Current existing software methodologies create inter-core overhead that results in low utilization of silicon and wasted cycles that increase power and latency. Our TruStream technology creates concurrency and scalability across a core fabric to solve major roadblocks for key vertical markets, delivering massive performance gains, with lower power usage, cost, latency and footprint.”
“Our intellectual property (IP) strategy is essential in building economic value for our company,” stated Yatish Mishra, president and COO of CORNAMI. “Beyond protecting inventions in this high-growth marketplace, we have strategically chosen our filings for their business value. Our IP assets have strong economic value because of their ability to create vastly greater efficiencies in processing than what is available today. The company will also see ROI from its products and services based on these inventions.”
Paul Master will be speaking at the upcoming SPARK SUMMIT (Boston, Feb 7-9) and showcasing CORNAMI’s success in delivering a significant performance increase for the Yahoo Streaming Benchmark. This standard benchmark measures real-time mobile advertising performance. CORNAMI supports Apache Spark, along with other big-data frameworks, to provide an easy-to-use environment for running existing applications.
CORNAMI™ is a high-performance computing (HPC) company that increases compute performance significantly, made possible by the development of its breakthrough and patented multi-core technology that efficiently uses heterogeneous cores in a highly concurrent, parallel manner. This technology will reduce the use of power sources and lower latency, while vastly increasing the compute performance of products and their ROI in a variety of markets, with the Company’s initial focus being Big Data specifically AI and Machine Learning. CORNAMI was founded by three Silicon Valley veterans of chip and reconfigurable computing technologies: Gordon Campbell, Dr. Fred Furtek and Paul Master. The Company’s headquarters is in Silicon Valley, with offices in Sacramento and Boston. www.CORNAMI.com