Dynamic Reception Buffer Allocation for SoC Throughput

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

In information processing devices, especially in high-performance systems-on-chip (SoCs) with limited mounting area, securing sufficient reception buffer size for multiple transmission sources is challenging, leading to throughput performance issues due to the difficulty in dynamically adjusting reception buffer sizes efficiently, especially when resources are limited.

Innovation Solution

The implementation of a shared reception buffer management system where transmission modules can dynamically request and lend buffer allocation based on usage, allowing flexible accommodation of the shared buffer area between modules to ensure throughput performance without the need for extensive buffer size allocation to each module individually.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If a reception buffer is allocated to each transmission module individually, then each module can ensure its throughput performance, but the total buffer size required becomes excessively large when multiple transmission sources are present

Engineering Contradiction:
Improvethroughput performanceVSAvoidreception buffer size
Core Design Contradiction:
ReliabilityVSQuantity of substance

Solution Approach 1:

Multiple transmission modules share a single common reception buffer instead of each having dedicated buffers. The buffer is dynamically allocated among transmitting modules based on current transmission needs, allowing efficient resource utilization while maintaining throughput performance for all modules

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

The reception buffer allocation is dynamically adjusted based on real-time transmission status. When a module's transmission queue exceeds a threshold, it can request additional buffer space from other modules, and the buffer allocation changes dynamically rather than being fixed

Inventive Principle:
Principle #15Dynamics

2Productivity

If the reception buffer size is increased to handle multiple transmission sources, then throughput performance improves, but the mounting area required in the system-on-chip increases

Engineering Contradiction:
Improvethroughput performanceVSAvoidmounting area
Core Design Contradiction:
ProductivityVSArea of stationary object

Solution Approach 1:

Instead of allocating separate buffer memory for each transmission module (which would multiply the total area), a single shared buffer is used that serves all modules. This consolidation dramatically reduces the total mounting area required while maintaining the ability to handle multiple transmission sources simultaneously

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

The buffer allocation parameters are dynamically adjusted based on transmission load. Modules can request buffer space when needed and release it when not needed, allowing the system to maintain high throughput performance only when actually required rather than reserving maximum buffer space continuously

Inventive Principle:
Principle #35Parameter changes

3Quantity of substance

If the reception buffer is shared among multiple transmission modules, then buffer resource efficiency improves, but the complexity of managing buffer allocation increases

Engineering Contradiction:
Improvebuffer resource efficiencyVSAvoidbuffer management complexity
Core Design Contradiction:
Quantity of substanceVSDevice complexity

Solution Approach 1:

Transmitting modules autonomously manage their own buffer needs by monitoring their transmission queues and making lending requests when required. The buffer management system automatically responds to these requests based on current allocation status, reducing the need for complex centralized control

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The system implements feedback mechanisms where transmission modules report their buffer usage status and receive responses about available buffer space. This feedback loop allows modules to self-regulate their buffer requests and maintain efficient resource utilization without complex external management

Inventive Principle:
Principle #23Feedback

Data Source

PatentUS11722428B2Information processing device and method of controlling information processing device
Publication Date: 2023.08.08 FUJITSU LTD
  • US11722428B2 patent drawing
  • US11722428B2 patent drawing
  • US11722428B2 patent drawing

AI summary

The reception module includes a reception buffer that is shared by the first and second transmission modules to store the packet. The first transmission module is configured to: transmit, to the second transmission module, a lending request of the reception buffer allocated to the second transmission module, based on a use amount of the reception buffer allocated to the first transmission module, and increase an allocation amount of the reception buffer for the first transmission module in a case where the second transmission module has transmitted an acceptance response to the lending request. The second transmission module is configured to: receive the lending request and transmit the acceptance response to the first transmission module based on a use amount of the reception buffer allocated to the second transmission module, and decrease an allocation amount of the reception buffer for the second transmission module when transmitting the acceptance response.