Chip Address Translation for High-Speed Data Transfer

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

The existing methods for increasing processing speed by dividing circuits into multiple chips and using high-speed serial interfaces like PCI Express require additional switches, leading to increased costs.

Innovation Solution

An information processing apparatus with multiple chips connected in series, where each chip includes a receiving unit, a register for address translation, a determination unit, an address translation unit, a controller unit, and a transmission unit, allowing for address translation and efficient data transfer between chips without the need for additional switches.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Speed

If PCI Express switches are added to connect multiple image processing units to the PCI Express interface, then high-speed data transfer between chips is achieved, but the cost is increased

Engineering Contradiction:
Improvedata transfer speedVSAvoidcost
Core Design Contradiction:
SpeedVSEase of manufacture

Solution Approach 1:

The patent introduces an address translation unit as an intermediary component within each chip to translate address information between different address spaces. This mediator enables direct peer-to-peer data transfer between chips without requiring external PCI Express switches, thereby achieving high-speed data transfer while reducing system cost by eliminating the need for additional switch hardware.

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent divides the address space into multiple segments, with each chip having its own address translation unit that manages local address space segmentation. This segmentation allows each chip to independently translate addresses to the destination chip's address space, enabling direct communication without centralized switching infrastructure and reducing overall system cost.

Inventive Principle:
Principle #1Segmentation

2Productivity

If circuits are divided into multiple chips to increase processing speed, then parallel processing capability is improved, but device complexity is increased

Engineering Contradiction:
Improveprocessing speedVSAvoidsystem complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent implements a universal address translation unit that can handle multiple functions: translating addresses to local chip memory, translating addresses to remote chips, and managing different address spaces. This multi-functional component simplifies the overall system architecture by providing a standardized mechanism for all inter-chip communications, reducing the complexity that would otherwise arise from having specialized communication paths for each function.

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

Solution Approach 2:

Each chip is equipped with its own address translation unit that autonomously translates address information without requiring external assistance from PCI Express switches or centralized controllers. This self-service capability allows each chip to independently manage its own address translation needs, simplifying the system architecture by eliminating the need for complex centralized address management infrastructure.

Inventive Principle:
Principle #25Self-service

Data Source

PatentUS10324866B2Information processing apparatus and data transfer method
Publication Date: 2019.06.18 CANON KK
  • US10324866B2 patent drawing
  • US10324866B2 patent drawing
  • US10324866B2 patent drawing

AI summary

An information processing apparatus includes a first, second, and third chips connected in series. The second chip includes a receiving unit, a register, a determination unit, an address translation unit, a controller unit, and a transmission unit. The receiving unit receives data and address information from the first chip. The determination unit determines whether the received address information corresponds to an address translation area based on address translation information set to the register. The address translation unit outputs translated address information to an internal bus. The controller unit controls to store data to which address information corresponding to an address area set for the second chip is attached. The transmission unit transmits to the third chip data to which address information is attached. The address translation unit translates address information corresponding to an address area set for the second chip into an address destination in the second chip.