Prefetch Metadata Storage Policy Switching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The limited storage capacity for metadata in prefetching systems can lead to reduced prefetch performance, as existing replacement policies may evict useful metadata entries before they can provide effective prefetching, especially when the number of active streams exceeds the capacity of the prefetch metadata storage.
Innovation Solution
Implementing an alternative replacement policy that locks recently used entries to prevent eviction, switching from a default 'least recently used' policy to a 'keep recently used' policy when the prefetch metadata storage capacity is exceeded, ensuring that valuable metadata entries remain allocated for longer periods and continue to provide effective prefetching.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If a default least recently used replacement policy is used for prefetch metadata storage, then the storage structure is simple and easy to implement, but useful metadata entries are evicted before they can provide effective prefetching when the number of active streams exceeds storage capacity
Solution Approach 1:
The system dynamically switches between two replacement policies based on runtime conditions. When the number of active streams exceeds the metadata storage capacity, the system transitions from a static least recently used policy to a dynamic alternative policy that locks recently used entries, preventing their eviction and maintaining prefetch performance under high stream conditions.
2Quantity of substance
If the prefetch metadata storage capacity is increased, then more metadata entries can be retained, but the device complexity and resource usage increase
Solution Approach 1:
The system changes the behavioral parameters of the replacement policy rather than increasing physical storage capacity. By modifying the eviction behavior through locking mechanisms and policy switching, the system maintains effective prefetching with limited storage resources, avoiding the need to expand the metadata storage structure.
3Duration of action of stationary object
If metadata entries are retained for longer periods, then prefetch accuracy improves, but the storage turnover and ability to adapt to new streams decreases
Solution Approach 1:
The system applies different retention characteristics to different metadata entries based on their usage patterns. Recently used entries are locked and retained longer to maintain prefetch accuracy, while other entries follow the standard replacement policy, allowing the system to adapt to new streams without compromising the effectiveness of active metadata entries.
Data Source
AI summary
Aspects of the present disclosure relate to an apparatus comprising processing circuitry, prefetch circuitry and prefetch metadata storage comprising a plurality of entries. Metadata items, each associated with a given stream of instructions, are stored in the prefetch metadata storage. Responsive to a given entry of the plurality of entries being associated with the given stream associated with a given metadata item, the given entry is updated. Responsive to no entry of the plurality of entries being associated with the given stream associated with a given metadata item, an entry is selected according to a default replacement policy, the given stream is allocated thereto, and the selected entry is updated based on the given metadata item. Responsive to a switch condition being met, the default selection policy is switched to an alternative selection policy comprising locking one or more entries by preventing allocation of streams to the locked entries.