Processing Performance Prediction via Virtual Resource Simulation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for predicting processing performance on new computers are inaccurate due to reliance on average usage time simulations, which differ from actual execution sequences, leading to insufficient prediction accuracy.
Innovation Solution
An information processing method that acquires execution information from existing computers, converts usage time based on resource ability values, and allocates resources on new computers to simulate task execution, allowing for accurate prediction of processing time by simulating resource management and idle times.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional average usage time simulation is used for performance prediction, then the prediction process is simple, but the prediction accuracy is insufficient
Solution Approach 1:
The patent creates a virtual copy of the actual computer system by extracting and simulating its resource configuration, task execution sequences, and resource allocation patterns. This virtual model replicates the behavioral characteristics of the physical system, enabling accurate performance prediction without requiring actual hardware testing. The copying principle is applied by creating a software-based simulation environment that mirrors the target system's operational dynamics.
Solution Approach 2:
The patent performs preliminary extraction and analysis of execution information from the actual computer system before conducting performance predictions. By pre-collecting resource usage data, task sequences, and allocation patterns, the system prepares a comprehensive dataset that enables accurate simulations. This preliminary action phase includes measuring actual resource usage, converting usage times based on resource ability values, and storing execution information for future prediction scenarios.
2Measurement precision
If detailed execution information and resource allocation simulation are implemented, then prediction accuracy improves, but computational complexity increases
Solution Approach 1:
The patent segments the complex simulation process into distinct functional modules: execution information extraction, resource ability value conversion, virtual run time calculation, and performance prediction. Each module handles a specific aspect of the simulation, processing task information, resource allocation, and time calculations separately. This segmentation reduces overall system complexity by making each component manageable and independently optimizable while maintaining comprehensive prediction capability.
Solution Approach 2:
The patent introduces execution information as an intermediary data structure that bridges the gap between actual system measurements and simulation requirements. This intermediary layer captures detailed resource usage patterns, task sequences, and allocation behaviors, serving as a mediator that translates physical system characteristics into simulation-ready formats. The execution information acts as a buffer that simplifies the interaction between data collection and prediction algorithms.
3Adaptability or versatility
If resource ability values and conversion formulas are used, then cross-computer prediction becomes possible, but calculation overhead increases
Solution Approach 1:
The patent transforms physical resource characteristics into standardized ability value parameters that enable cross-computer comparisons. By converting diverse hardware specifications (CPU speed, memory capacity, disk I/O rates) into dimensionless ability values, the system achieves universality in performance prediction. The conversion formulas establish mathematical relationships between different resource types and the standardized parameters, allowing consistent prediction across heterogeneous computer systems while optimizing calculation efficiency through parameter normalization.
Data Source
AI summary
An information processing method includes acquiring sets of execution information of a plurality of information processes executed by a first information processing apparatus, converting the usage time in each set of execution information into usage time on a second information processing apparatus, executing a resource allocation process of allocating the resource of the second information processing apparatus to a first information process during the converted usage time, allocating the resource of the second information processing apparatus to a second information process for idle time not allocated to the first information process during the converted usage time, and accumulating virtual run time of the allocated resources, and estimating execution time when executing the plurality of information processes on the second information processing apparatus on the basis of the accumulated virtual run time.