Prediction Circuitry for Storage Unit Access Latency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Determining which storage unit contains a requested data value in processing apparatuses can be time-consuming, slowing down performance, as existing methods require waiting for the target storage unit to be identified before initiating data access.
Innovation Solution
Incorporating prediction circuitry that detects specific types of data access instructions to predict the target storage unit based on instruction types, allowing data access to be initiated before the actual storage unit is identified, using techniques such as exploiting correlation between successive instructions and maintaining status flags for prediction accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the storage controller waits to identify the target storage unit before initiating data access, then the accuracy of data access is ensured, but the processing performance deteriorates due to increased latency
Solution Approach 1:
The prediction circuitry performs preliminary action by predicting the target storage unit before the storage controller finishes identifying it. The prediction is made based on the instruction type (e.g., load or store) and previous access patterns, allowing the data access to be initiated in advance without waiting for the complete address resolution, thus reducing latency while maintaining acceptable accuracy through statistical prediction models
2Loss of time
If prediction circuitry is added to predict the target storage unit, then processing latency is reduced, but device complexity increases
Solution Approach 1:
The prediction circuitry uses a simple, lightweight prediction mechanism that maintains a prediction buffer storing recent access patterns. Instead of implementing a complex deterministic identification system, the patent employs a statistically-based prediction approach that is computationally inexpensive and can be implemented with minimal additional circuitry, making the complexity increase acceptable given the significant latency reduction
3Use of energy by moving object
If prediction is made based on instruction type rather than address analysis, then power consumption is reduced, but prediction accuracy may decrease
Solution Approach 1:
The prediction circuitry extracts only the essential feature needed for prediction, which is the instruction type (load or store), and ignores the full address information for the prediction process. By taking out only the relevant instruction type characteristic and using it to index into the prediction buffer, the system achieves low-power operation while maintaining sufficient prediction accuracy for the given application context
Data Source
AI summary
A data processing apparatus comprises processing circuitry and a plurality of storage units. When the processing circuitry executes a data access instruction, then a storage controller identifies based on a target storage address of the data access instruction, which of the storage units includes the target storage location identified by the target storage address. Prediction circuitry is provided to predict a predicted storage unit predicted to include the target storage location, so that retrieval of the data value from the predicted storage unit can be initiated before the storage controller has identified the target storage unit. The prediction circuitry makes the prediction based on the type of the data access instruction executed by the processing circuitry.