Cross-Routed Communication Bridge for Multi-GPU Isolation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current data processing systems face challenges in providing sufficient independence and isolation between multiple graphics processing sub-systems, such as those found in vehicles, while also optimizing resource usage and cost, particularly in scenarios requiring multiple graphics processing units for different displays and advanced driver assistance systems.

Innovation Solution

A data processing system comprising multiple processing units connected via communications bridges, where each processing unit has an internal communications network routing communications in a specific sequence order, allowing functional units to use different connections based on their position relative to the communications bridge, enabling efficient communication and reducing bottlenecks between linked processing units.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If a single graphics processing unit is time shared between different graphics processing functions, then resource usage is optimized and cost is reduced, but independence and isolation between different sub-systems is insufficient

Engineering Contradiction:
Improveresource sharingVSAvoidindependence and isolation
Core Design Contradiction:
Adaptability or versatilityVSReliability

Solution Approach 1:

The system segments the graphics processing resources by providing multiple processing units (first GPU, second GPU) that can be independently allocated to different sub-systems. Each processing unit maintains its own functional units and communication paths, enabling both resource specialization and system independence simultaneously

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The communication bridge provides multi-functionality by enabling processing units to operate in multiple modes: independent operation for isolation-critical applications, and combined operation for resource sharing. The bridge can dynamically configure communication paths to support different operational scenarios

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

2Reliability

If a completely separate graphics processing unit is provided for each graphics processing function, then independence and isolation between sub-systems is achieved, but the number of processing components and cost increase

Engineering Contradiction:
Improveindependence and isolationVSAvoidnumber of processing components
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The system merges multiple processing units through the communication bridge to form an integrated multi-GPU system. This allows independent processing units to be combined when resource sharing is beneficial, reducing overall system complexity and cost while maintaining the option for independent operation when isolation is required

Inventive Principle:
Principle #5Merging (Combining)

3Productivity

If processing units are connected via communications bridge with multiple connections routing between different node positions, then communication efficiency is improved and bottlenecks are reduced, but device complexity increases

Engineering Contradiction:
Improvecommunication efficiencyVSAvoidcommunication routing configuration
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The communication bridge introduces an additional dimensional aspect to the communication architecture by creating cross-position connections between processing units. Instead of simple sequential routing, the bridge enables multi-path communication where data can flow through different node position combinations, adding routing dimensionality that eliminates bottlenecks

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

Data Source

PatentUS11604752B2System for cross-routed communication between functional units of multiple processing units
Publication Date: 2023.03.14 ARM LTD
  • US11604752B2 patent drawing
  • US11604752B2 patent drawing
  • US11604752B2 patent drawing

AI summary

A data processing system comprising a plurality of processing units. Each processing unit comprises a set of plural functional units and an internal communications network that routes communications between the functional units in a particular sequence order of the functional units. Each processing unit is connected to at least one other processing unit via a communications bridge that has at least two connections, a first connection that routes communications between a first pair of network nodes of the pair of processing units, and a separate, second connection that routes communications between a second, different pair of network nodes of the pair of processing units. Each connected pair of network nodes comprises network nodes having different positions in the internal communications network sequence order of the network nodes and/or network nodes associated with functional units of different types.