Ternary Content Addressable Memory Single-Cycle Write Mechanism

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current ternary content addressable memory (TCAM) architectures are insufficient in terms of write operation speed, which limits their applicability in high-demand network applications such as artificial intelligence, databases, and network switching.

Innovation Solution

The development of TCAM architectures that enable both data and mask writes to occur in a single clock cycle, utilizing control circuitry to generate control signals that allow for simultaneous writing of data and mask inputs across separate rows of memory cells, with buses dedicated to data and mask operations during write and search processes.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Speed

If data and mask are written in separate clock cycles, then the write operation is simpler to implement, but the write speed is insufficient for high-demand network applications

Engineering Contradiction:
Improvewrite operation speedVSAvoidcontrol circuitry complexity
Core Design Contradiction:
SpeedVSDevice complexity

Solution Approach 1:

The write operation is segmented into two independent but synchronized parts: data write controlled by first control signal and mask write controlled by second control signal. This segmentation allows parallel execution paths while maintaining independent control, enabling single-cycle completion without excessive complexity

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The control circuitry is pre-configured with edge-triggered logic that automatically generates the second control signal in response to the first control signal's edge. This preliminary setup eliminates the need for complex sequential control during operation, achieving fast writes with manageable complexity

Inventive Principle:
Principle #10Preliminary action

2Productivity

If separate buses are used for data and mask operations, then write and search operations can proceed independently, but the device complexity increases

Engineering Contradiction:
Improveoperational efficiencyVSAvoidbus structure complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The first bus and second bus are designed with dual functionality: during write operations, they independently transfer data and mask respectively; during search operations, they can be selectively activated. This multi-functionality allows independent data and mask writes while maintaining operational flexibility, achieving high productivity without excessive complexity

Inventive Principle:
Principle #6Universality (Multi-functionality)

Solution Approach 2:

The bus structure implements dynamic control where the first bus is activated for data operations and the second bus for mask operations during writes, while search operations can selectively use either bus based on timing edges. This dynamic activation pattern optimizes productivity by enabling parallel operations while managing complexity through conditional bus usage

Inventive Principle:
Principle #15Dynamics

Data Source

PatentUS11894054B2Methods for writing ternary content addressable memory devices
Publication Date: 2024.02.06 MEDIATEK SINGAPORE PTE LTD
  • US11894054B2 patent drawing
  • US11894054B2 patent drawing
  • US11894054B2 patent drawing

AI summary

Ternary content-addressable memory (TCAM) devices are described. The TCAMs described herein are designed to perform write operations—including data writes and mask writes—in a single clock cycle. For example, data input is written in a row of the TCAM during the first portion of a clock cycle, and a mask is written in another row of the TCAM during the second portion of the clock cycle, for example immediately after or after a programmable delay from the data write. In one implementation, a first bus is used both for data write and key search operations, and a second bus is used both for mask write and search masking operations. In another implementation, a first bus is used both for data write and key search operations, a second bus is used for mask write operations, and a third bus is used for search masking operations.