3D-NAND Search Engine for Dense In-Memory Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional TCAM technology for big data and AI processing suffers from low memory density and high power usage, while nonvolatile TCAM techniques require paired memory cells, limiting efficiency in searching and data storage.
Innovation Solution
A 3D search engine architecture using 3D-NAND memory with floating-gate transistors, enabling high-performance searching, comparing, and sorting by employing series-connected floating-gate transistors and a search encoder to drive word lines, and using sense amplifiers to generate match indications.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional TCAM technology is used for searching, then searching functionality is achieved, but memory density is low and power usage is high
Solution Approach 1:
The patent replaces conventional TCAM technology with 3D-NAND flash memory technology, substituting a different technical mechanism that achieves searching functionality through in-memory computing operations rather than traditional TCAM cell structures, thereby improving memory density while maintaining search capabilities
Solution Approach 2:
The patent changes the fundamental operating parameters by using 3D stacking architecture and floating gate transistor threshold voltage states to represent data, enabling high-density storage while performing search operations through voltage-based in-memory computing rather than conventional bit-cell approaches
2Reliability
If nonvolatile memory techniques are used for TCAM, then data retention is improved, but efficiency of searching and data storage is limited due to requiring paired memory cells
Solution Approach 1:
The patent extracts the pairing requirement constraint by using 3D-NAND flash memory's native single-cell structure for storing search keys, eliminating the need for paired memory cells and thereby improving searching efficiency while maintaining nonvolatile data retention characteristics
Solution Approach 2:
The patent makes the 3D-NAND flash memory cells universal by enabling them to perform both data storage and search key storage functions simultaneously, with the ability to dynamically configure which cells store search keys versus data, thereby improving overall system efficiency without requiring specialized paired structures
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
The 3D search engine achieves low latency, high content density, and robust reliability, supporting big data and AI applications with concurrent search operations and flexible matching capabilities.
Implementation Method 1
each memory string is enabled to couple the respective match and source lines that the respective memory string is coupled to via an a match-found impedance responsive to the configured states of the memory string exactly matching the control inputs of the series-connected floating-gate transistors of the memory string
Implementation Method 2
each sense amplifier is enabled to generate a respective indication of whether the match line it is coupled to is coupled to the source line via the match-found impedance
Implementation Method 3
Each of the configured states corresponds to a respective threshold voltage. Each of the respective control inputs is coupled to a respective gate of one more of the series-connected floating-gate transistors. The threshold voltages are settable via one or more programming operations.
Data Source
AI summary
A 3D search engine receives searches for application to word lines of a nonvolatile memory array. The engine uses two word lines per bit of information of the searches and two memory devices per bit of stored feature to search against, optionally enabling don't care and/or wildcard encoding. The engine uses respective bit lines of the nonvolatile memory array as respective matching lines for searching. Respective memory strings (e.g., NAND memory strings) of the nonvolatile memory array are usable to store respective data words, e.g., corresponding to features to search for. Respective pluralities of the memory strings are coupled in parallel to respective shared bit lines. Various encodings of features and searches enable exact, approximate, and range matching. The engine has applicability to comparing and sorting, in addition to searching in application areas such as artificial intelligence (AI) and big data.