Hybrid CPU GPU Task Management for Data Mining
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current data mining technologies face inefficiencies in processing graph-based data sets due to the heterogeneity of data distribution, leading to slower computations when using either CPUs or GPUs alone, as they are designed for uniform workloads and struggle with non-uniform data sets.
Innovation Solution
A hybrid task management system that dynamically allocates graph-based data mining tasks across multiple CPU and GPU cores, sorting tasks by uniformity and distributing non-uniform edges to CPUs and uniform edges to GPUs to minimize inter-processor communication and optimize workload balance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If tasks are allocated to GPU cores for data mining, then computation speed increases, but handling of non-uniform data sets becomes inefficient
Solution Approach 1:
The patent segments the data set into uniform and non-uniform portions, allocating uniform data to GPU cores for parallel processing and non-uniform data to CPU cores for sequential processing. This segmentation resolves the contradiction by allowing each processor type to handle the data portion it is best suited for, maintaining high speed while improving adaptability to non-uniform data.
Solution Approach 2:
The patent applies local quality by assigning different processing qualities to different data portions based on their uniformity characteristics. Uniform data receives GPU-based parallel processing with high throughput, while non-uniform data receives CPU-based sequential processing with higher adaptability. This local differentiation resolves the contradiction by optimizing processing quality for each data type.
2Adaptability or versatility
If tasks are allocated to CPU cores for data mining, then handling of non-uniform data improves, but computation speed decreases
Solution Approach 1:
The patent segments the data set into uniform and non-uniform portions, allocating uniform data to GPU cores for parallel processing and non-uniform data to CPU cores for sequential processing. This segmentation resolves the contradiction by allowing each processor type to handle the data portion it is best suited for, maintaining high speed while improving adaptability to non-uniform data.
Solution Approach 2:
The patent applies local quality by assigning different processing qualities to different data portions based on their uniformity characteristics. Uniform data receives GPU-based parallel processing with high throughput, while non-uniform data receives CPU-based sequential processing with higher adaptability. This local differentiation resolves the contradiction by optimizing processing quality for each data type.
3Productivity
If hybrid CPU-GPU task management is implemented, then overall data mining efficiency improves, but system complexity increases
Solution Approach 1:
The patent applies preliminary action by pre-classifying data elements into uniform and non-uniform categories before task allocation. This pre-processing step simplifies the hybrid system management by establishing clear allocation rules based on data characteristics, reducing the complexity of real-time decision-making while maintaining high productivity.
Solution Approach 2:
The patent introduces a task management system as an intermediary that automatically handles the allocation and coordination between CPU and GPU cores. This intermediary layer abstracts the complexity of hybrid processing, providing a unified interface for data mining operations while optimizing the underlying resource distribution to improve productivity.
Data Source
AI summary
Embodiments described herein provide a system for facilitating hybrid task management across a central processing unit (CPU) and a graphics processing unit (GPU) of a computer. During operation, the system determines a set of tasks for performing data mining on a data set and storing the set of tasks in a data structure in an ascending order of uniformity associated with a respective task. The uniformity of a task indicates how uneven and skewed the task is compared to other tasks in the set of tasks. The system then allocates a subset of tasks to a core of the CPU from a front of the data structure and a subset of tasks to a core of the GPU from a back of the data structure.