Per-group prefetch status to reduce duplicate requests

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Cache misses occur frequently due to inefficient prefetching techniques, leading to unnecessary data requests and increased traffic between the prefetch block and cache memory, as existing methods fail to effectively track and avoid duplicate prefetching of already prefetched data blocks.

Innovation Solution

A prefetch block detects access to data blocks and generates proposed blocks for prefetching, using prefetch tracking data to determine if a prefetch group has already been marked as prefetched, thereby avoiding redundant requests to the cache memory by not prefetching data that is already stored.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If prefetching techniques are used to reduce cache misses, then cache hit rate is improved, but duplicate prefetch requests increase traffic between prefetch block and cache memory

Engineering Contradiction:

Improvecache hit rateVSAvoidprefetch request volume

Core Design Contradiction:

ReliabilityVSQuantity of substance

Solution Approach 1:

The patent applies preliminary action by checking prefetch tracking data before issuing prefetch requests to determine if data has already been prefetched. This prevents duplicate prefetch requests by performing the verification action in advance, reducing unnecessary traffic between the prefetch block and cache memory while maintaining high cache hit rates

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent implements feedback through prefetch tracking data that records which data blocks have been prefetched. This feedback mechanism allows the prefetch block to monitor and control duplicate requests by comparing current prefetch candidates against the tracking data, thereby reducing redundant traffic while preserving cache effectiveness

Inventive Principle:

Principle #23Feedback

2Quantity of substance

If prefetch tracking data is implemented to avoid duplicate prefetching, then traffic between prefetch block and cache memory is reduced, but device complexity increases

Engineering Contradiction:

Improveprefetch request volumeVSAvoidprefetch tracking structure

Core Design Contradiction:

Quantity of substanceVSDevice complexity

Solution Approach 1:

The patent uses copying by maintaining a simplified tracking data structure that mirrors essential prefetch state information. Instead of complex tracking mechanisms, the system creates a lightweight copy of prefetch status data that can be efficiently checked, reducing traffic volume while adding minimal complexity to the prefetch block structure

Inventive Principle:

Principle #26Copying

Data Source

PatentUS20200167287A1Per-group prefetch status to reduce duplicate prefetch requests

Publication Date: 2020.05.28 ATI TECHNOLOGIES ULC

AI summary

A technique for prefetching data for a cache is provided. The technique includes detecting access to a data block. In response to the detection, a prefetch block generates proposed blocks for prefetch. The prefetch block also examines prefetch tracking data to determine whether a prefetch group including the proposed blocks is marked as already having been prefetched. If the group has been marked as already having been prefetched, then then prefetch block does not prefetch that data, thereby avoiding traffic between the prefetch block and the cache memory. Using this technique, unnecessary requests to prefetch data into the cache memory are avoided.