Cache Persistence Policy for Non-Temporal Workload Thrashing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Cache thrashing occurs when workloads stream large amounts of data from memory, leading to frequent cache misses, evictions, and inefficient use of system resources, which negatively impact processor performance and efficiency.
Innovation Solution
Monitor cache activity to identify non-temporal workloads and alter the persistence policy by reducing the tenure or directly casting out cache lines associated with these workloads, using metrics such as cache hit/miss ratio and fetch/cast-out ratio to minimize cache thrashing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If data is persisted in cache for streaming workloads, then cache capacity is utilized, but cache misses and evictions increase due to non-referenced data occupying cache space
Solution Approach 1:
The patent changes the persistence parameter (tenure) of cache lines based on workload type. For non-temporal workloads, the persistence tenure is reduced or disabled, allowing cache lines to be evicted after a single access. This parameter change prevents non-referenced data from occupying cache space indefinitely, resolving the contradiction between cache utilization and processor performance
Solution Approach 2:
The patent dynamically adjusts the persistence policy based on detected workload characteristics. When non-temporal workload patterns are identified (such as streaming workloads), the system automatically modifies cache line tenure settings. This dynamic adaptation allows the cache to optimize between temporal and non-temporal workloads, preventing performance degradation while maintaining capacity utilization
2Reliability
If cache line tenure is extended for all workloads, then cache hit ratio improves, but system resources are wasted on persisting non-referenced data
Solution Approach 1:
The patent applies different persistence qualities to different cache lines based on their associated workload type. Temporal workload cache lines receive extended tenure for high hit ratios, while non-temporal workload cache lines receive reduced or disabled persistence. This localized quality differentiation resolves the contradiction by applying resource-intensive persistence only where it provides benefit
Solution Approach 2:
The system changes the persistence parameter (tenure) based on workload classification. For non-temporal workloads, the tenure parameter is reduced or disabled, preventing waste of system resources on data that will not be re-referenced. This selective parameter adjustment maintains cache hit ratios for beneficial data while eliminating resource waste
3Stability of the object's composition
If traditional cache persistence policy is used for streaming workloads, then cache structure is maintained, but cache thrashing occurs with frequent misses and evictions
Solution Approach 1:
The patent introduces dynamic persistence policy adjustment that responds to workload characteristics. When streaming or non-temporal workloads are detected, the system dynamically modifies cache line tenure settings to prevent thrashing. This dynamic behavior allows the cache to maintain structural stability for temporal workloads while adapting to prevent performance degradation in non-temporal scenarios
Solution Approach 2:
The system changes the persistence tenure parameter based on detected workload patterns. For non-temporal workloads exhibiting streaming behavior, the tenure parameter is reduced or disabled, preventing the cache from attempting to persist data that will not be re-referenced. This parameter change eliminates cache thrashing while maintaining cache structure for workloads where persistence is beneficial
Data Source
AI summary
implemented method, system, and computer program product for minimizing the effects of cache thrashing involving non-temporal workloads. The cache activities of a workload, including the cache activities (e.g., number of cache hits) involving local and peer caches, are monitored. Based on analyzing the metrics of such monitored cache activities, a determination is made as to whether a non-temporal workload is identified. For example, such a determination may be based on comparing the metrics of the monitored cache activities of the workload to a threshold value. Upon identifying a non-temporal workload, the cache line(s) associated with the non-temporal workload are identified. The persistence policy for the identified cache line(s) is then altered. For example, the persistence policy for the identified cache line(s) may be altered by reducing the tenure of such a cache line(s) thereby reducing the number of cache misses or evictions and minimizing the effects of cache thrashing.