Storage Module Data Comparison for Read-Write Conflict Verification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The challenge in verifying storage modules with independent read and write channels using UVM verification platforms is the uncertainty in read data due to potential read-write conflicts, leading to inaccurate abnormality determinations.
Innovation Solution
A method involving monitoring read and write transactions, storing them in separate queues, and using a scoreboard to compare data values considering potential conflicts, ensuring accurate data comparisons by using alternative values from the write transaction queue.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If a UVM verification platform is used to verify the storage module with independent read and write channels, then the verification process can be automated, but the accuracy of abnormality determination is lowered due to read-write conflicts causing uncertainty in monitored read data
Solution Approach 1:
The patent applies preliminary action by establishing a write transaction queue in advance to store write transactions that may conflict with read transactions. Before the data comparison occurs, the system pre-preares the write transaction information (including original values before writing) so that when a read transaction is monitored, the corresponding write transaction data is already available for comparison, thereby resolving the uncertainty of read data in read-write conflict scenarios.
Solution Approach 2:
The patent introduces an intermediary mechanism by creating a write transaction queue that acts as a mediator between the read transaction monitoring and the actual data comparison. This queue stores write transaction information as an intermediate data structure, allowing the scoreboard to access and compare write data values (both current and original) when evaluating read transactions, thus mediating the complexity of read-write conflict resolution.
2Measurement precision
If the reference model is designed to accurately simulate read-write conflict processing characteristics, then the measurement precision of abnormality determination is improved, but the device complexity increases significantly
Solution Approach 1:
The patent applies segmentation by dividing the verification functionality into separate components: a monitor for collecting read and write transactions, a write transaction queue for storing write transaction information, and a scoreboard for performing data comparison. This segmentation allows each component to have a specific, manageable function, avoiding the need for a single complex reference model while achieving accurate read-write conflict resolution.
Solution Approach 2:
The patent uses copying by creating a simplified representation of write transaction information in the write transaction queue. Instead of requiring the reference model to fully simulate complex read-write conflict behavior, the system copies essential write transaction data (address, original value, new value) into a dedicated queue structure, which then provides the necessary comparison data to the scoreboard, thereby simplifying the overall reference model.
Data Source
AI summary
The present application relates to data comparison of a storage module. A method includes monitoring read transactions and write transactions, storing a write transaction that is in-transit and a write transaction that has ended but may have a read-write conflict with a read transaction, storing a read transaction that is in-transit and a read transaction that has ended but has not completed data comparison, and performing a data comparison on each ended read transaction by using a scoreboard. When a data comparison between a read data value in the ended read transaction and a value at a corresponding address in a reference model fails, and a data comparison between the read data value and a write data value and/or an original value of one write transaction passes, a data comparison on the ended read transaction passes. This method may improve an accuracy of determining an abnormity of a storage module.