Skip to main content


Moore’s Law and Von Neumann

Beginning in the early 2000’s, computer-processor design stalled out. The semiconductor industry could no longer increase the speed of new processors in order to increase software performance because there was simply no way to cool them. By 2004, semiconductor industry design, led by Intel, added multiple processors (cores) on a single chip so that they could continue to support Moore’s Law that is, in essence, the doubling microprocessor performance every two years. The common thinking was that advances in performance would come from adding additional processing cores rather than by increasing speed.

Unfortunately, with few exceptions, the world’s 25 million plus programmers do not have a good way to program and fully utilize multiple cores. Nearly all the world’s applications typically use only a single core at any given time, and the unused cores remain dark (idle).


Why is this? Because the von Neumann programming model, which underlies modern computing, works brilliantly when dealing with a single core, but fails spectacularly when dealing with multiple cores. Simply put, the mechanisms we are using today for multi-core programming are inherently inadequate, and the complexities required to deal with multiple cores far exceeds the limits of most programmers. The result is inefficient systems with large amounts of untapped compute capacity remaining idle. The biggest problem facing higher levels of compute performance is now a HUGE opportunity for CORNAMI.

CORNAMI has developed and patented new concurrency technology that uniquely changes software performance, power use, latency and platform footprint. By using a standardized runtime concurrency model called TruStream, heterogeneous multi-core processor resources are abstracted into a common homogenous core pool. Programmers can easily implement concurrency through CORNAMI’s TruStream control structures embedded in standard languages.

TruStream is implemented in software and runs on single or networked heterogeneous multi-core CPUs and operating systems (x86, ARM, Android, Linux, and Mac OS).


Take High Performance Computing to the Next Level

TruStream Concurrency

CORNAMI’s TruStream implements a highly efficient and extensible model of concurrent programming solving the hardest problem in computer science today. TruStream supports a direct concurrency abstraction, allowing applications to seamlessly scale using a fabric of cores ranging from intra-CPU to WAN-linked datacenters. vLanguage, CPU type, and o/s agnostic, TruStream supports concurrency simply and deterministically (no locks) increasing performance, with no performance falloff, as more cores are utilized. Whether it be purely with software or with hardware augmentation, TruStream increases performance, reduces power consumption and latency, and introduces additional forms of application concurrency previously unavailable.

TruStream Core Fabric TSCF  & DCIC Data Center on CHIP

CORNAMI’S unique technology can take a topology with interconnections and actions and efficiently accelerate it utilizing its “TruStream computing as a fabric” (TSCF) on its ultra-high core density, ultra-high memory bandwidth data center chip (DCIC).

CORNAMI’s intelligent architecture can scale software to an almost endless “sea of cores” operating concurrently with high speed inter-connects and programmability without the overhead of inefficient communication between today’s multi-core computing systems.  This innovative approach increases performance an order of magnitude and more due to efficient utilization, which reduces latency and power significantly.

CORNAMI will reset the current approach to data center footprints based on compute performance to power, reduced latency and overall cost by accelerating Big Data frameworks through an entirely new architecture that can scale almost infinitely across cores, systems, cloud and clusters.   All while reducing latency and power slashing operational costs by as much as 90%.

Additionally, CORNAMI’s chip architecture is programmable and “source compatible” unlike fixed function silicon in today’s GPU and ASICs, providing reduced obsolescence of hardware with built in flexibility to scale with company’s needs and requirements.

  • Patented TruStream parallelism technology built into each core for near linear scaling in performance
  • TruStream programming SDK model to utilize all cores efficiently from a higher level language such as C/C++
  • Programmable delivering future-proof products
  • Same chip can be used in any applications such as Big Data, Machine Learning, High Frequency Trading, Genomics, etc.
  • Different from typical FPGA single-function implementations or limited GPU accelerated routines
  • Ideal for Scale-UP and Scale-Out applications

FPGAs – field programmable gate arrays allow for continuing updates through software at the hardware level to add and/or upgrade feature sets enabling future requests and requirements. FPGAs are not constrained by bus widths, provide much lower latency than fixed function CPU/GPU architectures, and offer true parallelism at the hardware level.

Spectacular savings in data center construction and operations costs compared to today’s traditional IT Infrastructure

  • Powerful –Process and analyze data to drive competitive advantage
  • Efficient – Compute and data to improve operational efficiency
  • Simple – easy to deploy and operate
  • Compatible – leverage your existing investments for rapid adoption and low TCO
  • Economical – up to 90% more cost effective than traditional solutions