Slotted Ring Interconnect for Snoop Response Aggregation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

In data processing systems with multiple caching nodes, the existing ring-based interconnect structures face challenges in reducing snoop response traffic, which is essential for maintaining data coherency and reducing power consumption.

Innovation Solution

The implementation of a slotted ring architecture that allows snoop requests to identify specific slots for snoop responses, enabling aggregation and merging of responses within these slots as they circulate around the ring, thereby reducing the overall response traffic.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If a ring-based interconnect structure is used for coherency management, then coherency between multiple caching nodes is maintained, but snoop response traffic increases significantly

Engineering Contradiction:
Improvedata coherencyVSAvoidsnoop response traffic
Core Design Contradiction:
ReliabilityVSQuantity of substance

Solution Approach 1:

Multiple individual snoop response signals from different caching nodes are merged into a single aggregated snoop response signal. The interconnect node combines responses by OR-ing the signal states, so that multiple nodes can respond simultaneously without creating separate traffic streams for each node, thereby reducing overall traffic quantity while maintaining coherency information from all nodes.

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

The interconnect node serves multiple functions: it receives snoop requests from the coherency manager, distributes them to relevant caching nodes, collects individual responses, merges them into an aggregated response, and forwards the combined signal back to the coherency manager. This multi-functional approach eliminates the need for separate dedicated response paths for each caching node.

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

2Reliability

If separate snoop response paths are provided for each caching node, then response delivery is ensured, but wiring requirements and device complexity increase

Engineering Contradiction:
Improveresponse deliveryVSAvoidwiring requirements
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

Multiple separate response paths from individual caching nodes are merged into a single shared response path. The interconnect node aggregates responses from multiple nodes onto one signal line, reducing the number of wires required in the interconnect structure while ensuring that coherency information from all nodes can still be delivered to the coherency manager.

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

The interconnect node performs multiple functions including request distribution, response collection, signal merging, and data forwarding, thereby reducing the need for separate dedicated wiring paths for each caching node and simplifying the overall interconnect architecture.

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

3Quantity of substance

If snoop responses are aggregated in slots circulating around the ring, then response traffic is reduced, but response delivery timing may be delayed

Engineering Contradiction:
Improveresponse trafficVSAvoidresponse delivery timing
Core Design Contradiction:
Quantity of substanceVSLoss of time

Solution Approach 1:

The system uses periodic slot circulation around the ring, where slots are reused in a sequential manner. Snoop responses are placed in slots that circulate periodically, allowing multiple responses to be aggregated in the same slot across different cycles. This periodic mechanism reduces traffic by reusing slots while maintaining predictable timing through the regular circulation pattern.

Inventive Principle:
Principle #19Periodic action

Solution Approach 2:

The coherency manager issues snoop requests in advance with designated slot identifiers, and the interconnect node prepares to aggregate responses in the predetermined slot. This preliminary coordination ensures that responses are ready when the slot arrives at the interconnect node, minimizing delays while enabling traffic reduction through slot reuse.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentUS8635411B2Data processing apparatus and method for managing coherency of cached data
Publication Date: 2014.01.21 ARM LTD
  • US8635411B2 patent drawing
  • US8635411B2 patent drawing
  • US8635411B2 patent drawing

AI summary

An interconnect having a plurality of interconnect nodes arranged to provide at least one ring, a plurality of caching nodes for caching data coupled into the interconnect via an associated one of said interconnect nodes, and at least one coherency management node for implementing a coherency protocol to manage coherency of the data cached by each of said caching nodes. Each coherency management node being coupled into the interconnect via an associated one of said interconnect nodes. When each caching node produces a snoop response for said snoop request, the associated interconnect node is configured to output that snoop response in one of said at least one identified slots. Further, each interconnect node associated with a caching node has merging circuitry configured, when outputting the snoop response in an identified slot, to merge that snoop response with any current snoop response information held in that slot.