18091441. DATA DEPENDENCY-AWARE SCHEDULING simplified abstract (Advanced Micro Devices, Inc.)
Contents
DATA DEPENDENCY-AWARE SCHEDULING
Organization Name
Inventor(s)
Harris Gasparakis of Santa Clara CA (US)
DATA DEPENDENCY-AWARE SCHEDULING - A simplified explanation of the abstract
This abstract first appeared for US patent application 18091441 titled 'DATA DEPENDENCY-AWARE SCHEDULING
Simplified Explanation:
The patent application describes a processing system that efficiently schedules workgroups across kernels based on data dependencies to improve processing efficiency. Workgroups are divided into subsets based on data dependencies, with workgroups that produce data scheduled to run before workgroups that consume that data. This system optimizes data access by limiting subset sizes to fit local caches, reducing the need for memory access.
- Workgroups are scheduled across kernels based on data dependencies.
- Workgroups are divided into subsets based on data dependencies.
- Workgroups that produce data are scheduled before those that consume it.
- Subset sizes are limited to fit local caches, reducing memory access needs.
Potential Applications: This technology could be applied in various fields such as high-performance computing, data processing, and parallel computing systems.
Problems Solved: - Efficient scheduling of workgroups across kernels. - Optimized data access by reducing memory access needs.
Benefits: - Improved processing efficiency. - Enhanced data access optimization. - Increased performance in parallel computing systems.
Commercial Applications: Title: Enhanced Workgroup Scheduling System for High-Performance Computing This technology could be utilized in industries requiring high-speed data processing, such as finance, scientific research, and artificial intelligence.
Prior Art: Researchers can explore prior art related to parallel computing systems, data dependencies, and workgroup scheduling algorithms.
Frequently Updated Research: Stay updated on the latest advancements in parallel computing systems, data scheduling techniques, and optimization algorithms.
Questions about Workgroup Scheduling Technology: 1. How does the system determine the order of workgroup execution based on data dependencies? 2. What are the potential limitations of scheduling workgroups across kernels in parallel computing systems?
Original Abstract Submitted
A processing system flexibly schedules workgroups across kernels based on data dependencies between workgroups to enhance processing efficiency. The workgroups are partitioned into subsets based on the data dependencies and workgroups of a first subset that produces data are scheduled to execute immediately before workgroups of a second subset that consumes the data generated by the first subset. Thus, the processing system does not execute one kernel at a time, but instead schedules workgroups across kernels based on data dependencies across kernels. By limiting the sizes of the subsets to the amount of data that can be stored at local caches, the processing system increases the probability that data to be consumed by workgroups of a subset will be resident in a local cache and will not require a memory access.
