Enabling Technology

The Cell Structured Computer Chip (CSCC) Technology

The CSCC is a unique chip design applying fractal topology with sophisticated timing and sense networks. It enables real-time artificial intelligence with extensive processing power using precisely timed operations.

  • The CSCC is a fractal based computer chip specifically designed for real-time artificial intelligence.
  • CSCC consists of configurable logistic cells of RGUs and time-driven neural networks in a fractal topology.
  • All configurable logistic cells are fully connected by a switch fabric.
  • The fractal structure is ideal for recurrent deep learning by grouping GRUs in a local domain for an efficient cognitive process.

The Super Linear Concurrent Architecture

A single-layer unified cache system with a protocol for all pipelines that uses a real-time AI powered predictive MMU looking ahead at the data needed by all pipelines and then preparing it in advance.

  • 50% increase in concurrent memory access efficiency using real-time AI to reducing data exceptions in a multicore CPU
  • High-level simultaneous parallel processing for use in real-time AI functions
  • 10x increase in processor efficiency compared to existing technology

Real-Time Cognitive Neural Network by CSCC


CSCC:A Real-Time CNN

The CSCC is a unique chip design applying fractal topology with sophisticated timing and sense networks. It enables real-time artificial intelligence with extensive processing power using precisely timed operations.

  • A Cognitive Neural Network (CNN) is a network that immediately learns and adapts during real-time operation
  • The Cell Structured Computer Chip (CSCC) is a fractal based computer chip specifically designed for real-time artificial intelligence
  • CSCC consists of configurable logistic cells of RGUs and time-driven neural networks in a fractal topology
  • All configurable logistic cells are fully connected by a switch fabric

A Fractal Structure Example

Deep Learning using CSCC

  • All configurable logistic cells are fully connected by a switch fabric to form a complete network. The cells have enormous calculating power for challenging tasks.
  • There is a link between every cell that is controlled by a programmable switch. The links create a connected network that behaves like a neural network.
  • There are three independent CNN layers in the CSCC
    • Sense Network for cognition
    • Timing Network for real-time processing
    • Data Network for non-linear operations
  • The fractal structure is ideal for recurrent deep learning by grouping GRUs in a local domain for an efficient cognitive process.

Non-Linearity Channel Processing in CSCC

  • CSCC uses Channel Processing to handle non-linear operations in the fractal topology
  • The Channel Processing architecture, which consists of GRU cells and Control Pipelines, enables data to be processed while moving from one cell to another
  • Control Pipelines can process data from several GRU cells during a singe clock interval
  • The GRU cells maintain data coherency through Control Pipelines, enabling the highest parallelism
  • Data is no longer required to travel through the whole system to be processed by a central unit