Parallel Simulation Restore Points for Multicore Deadlock Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional simulation methods are inadequate for multicore processors, multiprocessors, and multi-master/slave devices as they struggle to efficiently handle deadlocks, I/O races, and multiple access to shared variables during parallel processing, which are common in these architectures.
Innovation Solution
A simulation method that sets restore points in each core model for parallel processing and stores information to reproduce the state of each core model at these points, allowing for the detection of issues like deadlocks and I/O races, and optimizing timing performance in parallel programs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional simulation methods are used for multicore processors, then single-core simulation can be performed, but parallel processing issues such as deadlocks, I/O races, and multiple access to shared variables cannot be efficiently detected
Solution Approach 1:
The patent divides the simulation system into separate core models (PE0, PE1, etc.), each with its own instruction set simulator and restore point management. This segmentation allows independent tracking of each core's execution state while maintaining the ability to detect inter-core issues like deadlocks and I/O races through coordinated restore point setting across multiple cores.
2Quantity of substance
If restore points are set in shared surrounding devices only, then memory consumption is reduced, but states of multiple core models cannot be reproduced simultaneously
Solution Approach 1:
The patent segments restore point storage by allocating separate restore points (A0, B0, C0 for PE0; A1, B1, C1 for PE1) to each core model. This allows simultaneous reproduction of multiple core states without requiring excessive shared memory, as each core's state is independently managed while still enabling detection of parallel processing issues.
Solution Approach 2:
The patent extends the restore point concept from a single shared resource to a multi-dimensional structure where each core model has its own set of restore points. This dimensional expansion enables simultaneous state reproduction for multiple cores without proportionally increasing memory consumption, as the restore points are organized in a structured manner across different core dimensions.
3Device complexity
If single thread simulation is used, then device complexity is reduced, but timing performance analysis of parallel programs cannot be optimized
Solution Approach 1:
The patent creates a universal simulation framework that can handle both single-thread and parallel processing simulations through the same core model structure. Each core model possesses multi-functionality, capable of executing instructions independently while participating in coordinated restore point setting, enabling timing performance analysis of parallel programs without significantly increasing device complexity.
Data Source
AI summary
A simulation method and apparatus including a restore point setting unit setting restore points in core models for executing threads using parallel processing. The method also includes storing information for reproducing a state the core models at the restore points.