On-Chip Code Breakpoint Debugging With Real-Time Off-Chip Memory

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Traditional chip algorithm debugging is inefficient due to the inability to output results in real time, requiring data transfer to another platform, leading to high bandwidth requirements and energy consumption.

Innovation Solution

Implementing an operation device with operation modules that store required data in their own storage units, reducing data access bandwidth and enabling parallel computational operations, thereby eliminating the need for large storage devices and enhancing scalability.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Speed

If data required by each operation module is stored in a memory, then data access bandwidth is increased, but device cost and bandwidth requirements increase

Engineering Contradiction:
Improvedata transfer speedVSAvoidbandwidth requirements
Core Design Contradiction:
SpeedVSQuantity of substance

Solution Approach 1:

The patent divides the centralized memory storage into distributed storage units, where each operation module has its own local storage unit. This segmentation reduces the bandwidth requirements of the centralized memory while maintaining fast data access for each operation module.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent implements local storage units within each operation module to store data locally, reducing the need for high-bandwidth centralized memory access. This local quality approach allows each module to access its required data quickly without competing for centralized memory bandwidth.

Inventive Principle:
Principle #3Local quality

2Productivity

If a large storage device with high data read and write speed is used, then data access performance is improved, but device cost increases

Engineering Contradiction:
Improvedata access performanceVSAvoiddevice cost
Core Design Contradiction:
ProductivityVSEase of manufacture

Solution Approach 1:

The patent segments the storage function into multiple small storage units distributed across operation modules rather than using one large high-performance storage device. This approach achieves high data access performance for each module while using lower-cost, smaller storage units.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

Each operation module serves its own data storage needs through its local storage unit, eliminating the need for a centralized high-performance storage device. This self-service approach reduces overall device cost while maintaining adequate data access performance.

Inventive Principle:
Principle #25Self-service

3Measurement precision

If data is copied to another platform for output, then complete chip algorithm results are obtained, but debugging efficiency decreases

Engineering Contradiction:
Improvealgorithm result accuracyVSAvoiddebugging efficiency
Core Design Contradiction:
Measurement precisionVSProductivity

Solution Approach 1:

The patent introduces a data output unit as an intermediary that enables real-time data output from the chip during debugging. This intermediary component allows debuggers to observe intermediate results without copying data to another platform, significantly improving debugging efficiency while maintaining result accuracy.

Inventive Principle:
Principle #24Intermediary (Mediator)

4Adaptability or versatility

If operation modules access centralized memory, then data sharing is facilitated, but bandwidth requirements and energy consumption increase

Engineering Contradiction:
Improvedata sharing capabilityVSAvoidenergy consumption
Core Design Contradiction:
Adaptability or versatilityVSUse of energy by moving object

Solution Approach 1:

The patent segments the centralized memory into distributed storage units associated with each operation module. This segmentation reduces the need for high-bandwidth centralized memory access, thereby reducing energy consumption while maintaining data sharing capabilities through the interconnection network.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent changes the storage architecture parameter from centralized to distributed, which reduces the bandwidth and energy requirements for data access while maintaining the ability of operation modules to share data when needed.

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentEP3825841B1On-chip code breakpoint debugging method, on-chip processor, and chip breakpoint debugging system
Publication Date: 2025.08.06 SHANGHAI CAMBRICON INFORMATION TECH CO LTD
  • EP3825841B1 patent drawingFigure 3-1
  • EP3825841B1 patent drawingFigure 3-2~3-3
  • EP3825841B1 patent drawingFigure 3-4~3-5

AI summary

The present invention relates to an on-chip code breakpoint debugging method, an on-chip processor, and a chip breakpoint debugging system. The method comprises: the on-chip processor starts and executes an on-chip code, and an output function is set at a breakpoint position of the on-chip code; the on-chip processor obtains output information of the output function, the output information is output information of the output function when the on-chip code is executed to the output function; the on-chip processor stores the output information into an off-chip memory. In the embodiments of the present invention, according to the output information, which is stored in the off-chip memory, of the output function, the on-chip processor can obtain execution conditions of breakpoints of the on-chip code in real time, can achieve the purpose of debugging multiple breakpoints in the on-chip code at the same time, and debugging efficiency of the on-chip code is improved.