Sections
Access Technologies
Image Slideshow
Right Column
Text Area

Digital Compute-In-Memory (DCIM) Enabled Portable Genome Analysis Prototype

Text Area

Description of Invention

Digital Compute-In-Memory (DCIM) Enabled Portable Genome Analysis Prototype
Image Slideshow
Finding Consensus Software Architecture
Image Slideshow
Manual Workflow vs AutoDCIM Workflow
Image Slideshow
DCIM-based Hardware Architecture
Image Slideshow
Text Area

Digital Compute-In-Memory (DCIM), revolutionizes AI chips by physically merging memory and computing units, drastically improving energy efficiency and processing speed. To enable the widespread adoption of this cutting-edge technology, we developed AutoDCIM, the first-ever Electronic Design Automation (EDA) tool tailored for DCIM design. AutoDCIM streamlines the DCIM circuits development process, making it faster and easier to implement DCIM-based solutions across various industries. Our invention, a portable genome analysis prototype, highlights the real-world potential of DCIM technology. This prototype enables faster, more cost-effective genome analysis, making advanced healthcare tools accessible even in remote or resource-constrained environments. By addressing the limitations of traditional computing platforms, our invention supports industries and communities with smarter, greener, and more efficient solutions, opening new possibilities for AI-driven applications, healthcare, and beyond.

Text Area

Key Technology Edges

  • The first automateod DCIM chip generation tool for specific applications, AutoDCIM.
  • Poliformer —— The first Transformer-based finding consensus DNA software architecture.
  • The first DCIM-based finding consensus DNA prototype, which provides 5.61x and 301.6x throughput speedup and energy saving than a current AI chip.