Link Disconnect Mechanism for Split Topology Communication Fabric
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The complexity of systems on chips (SoCs) with split topologies in communication fabrics complicates efficient link disconnection and powering down, as logical binding between data and control links cannot be entirely removed due to performance and functional constraints, leading to increased power consumption.
Innovation Solution
Implementing a method where link controllers in clusters determine idle resources and send separate requests for link disconnect via upstream control and data links, requiring dual acknowledgments before proceeding with disconnection, ensuring both data and control links are idle before powering down, thus maintaining connection and preventing premature link disconnection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of energy
If link disconnect sequences are implemented in a split topology communication fabric, then power consumption is reduced, but device complexity increases due to the need to manage separate data and control links
Solution Approach 1:
The link disconnect mechanism is segmented into separate handling for control links and data links. The link controller independently manages disconnect sequences for each link type, allowing precise control over when each link is disconnected based on its own idle status, thereby reducing overall power consumption while maintaining manageable complexity through modular management.
Solution Approach 2:
The link disconnect mechanism employs dynamic state monitoring where the link controller continuously assesses the idle status of both control and data links. Disconnect decisions are made dynamically based on real-time conditions, with the controller adapting the disconnect sequence timing and progression to match the actual operational state of the communication fabric.
2Reliability
If logical binding between data links and control links is maintained for performance reasons, then link disconnect efficiency is reduced, but functional reliability is improved
Solution Approach 1:
The disconnect process is segmented into independent phases for control links and data links. The link controller can evaluate and initiate disconnect sequences for each link type separately, allowing the system to maintain the necessary logical binding for reliability while executing disconnect operations efficiently without requiring complete logical coupling.
Solution Approach 2:
The link controller performs preliminary assessment of idle status for both control and data links before initiating disconnect sequences. This preliminary action allows the system to prepare for efficient disconnection while maintaining functional reliability through advance verification that disconnect conditions are met for both link types.
3Use of energy by stationary object
If separate disconnect requests are sent for control and data links, then power management precision is improved, but communication overhead increases
Solution Approach 1:
The link controller merges the monitoring and decision-making functions for both control and data links into a single coordinated process. While separate disconnect requests are generated for precision power management, the controller consolidates the assessment logic and timing coordination, reducing the overall communication overhead by eliminating redundant status checks and centralized coordination messages.
Data Source
AI summary
A system and method for efficiently disconnecting links in a communication fabric. A computing system includes a fabric for routing traffic among one or more agents and a memory controller for system memory. The fabric includes multiple hierarchical clusters with a split topology where the data links are physically separated from the control links. When the controller in an intermediate cluster or agent determines local resources are idle, the controller sends separate requests for link disconnect on each of the upstream control and data links. When a single acknowledgment of link disconnect is received, the controller does not proceed with any link disconnect. Rather, the controller waits until an acknowledgment of link disconnect is received for each of the upstream control and data links. If local resources remain idle, the controller sends confirmation upstream and performs a link disconnect for each of the upstream control and data links.