18091443. DYNAMIC CONTROL OF WORK SCHEDULING simplified abstract (Advanced Micro Devices, Inc.)
Contents
DYNAMIC CONTROL OF WORK SCHEDULING
Organization Name
Inventor(s)
Suchita Pati of Santa Clara CA (US)
Shaizeen Aga of Santa Clara CA (US)
Nuwan Jayasena of Santa Clara CA (US)
Matthew David Sinclair of Bellevue WA (US)
DYNAMIC CONTROL OF WORK SCHEDULING - A simplified explanation of the abstract
This abstract first appeared for US patent application 18091443 titled 'DYNAMIC CONTROL OF WORK SCHEDULING
Simplified Explanation:
The patent application describes a processing system with a scheduling mechanism that enables fine-grained reordering of workgroups in a kernel to facilitate communication across devices and overlap computations with network communication.
Key Features and Innovation:
- Scheduling mechanism for reordering workgroups of a kernel.
- Enables overlapping of producer computation with network communication.
- Facilitates data transmission between parallel processors in desired traffic patterns.
Potential Applications: This technology could be used in:
- High-performance computing systems.
- Distributed computing environments.
- Data-intensive applications requiring efficient communication.
Problems Solved:
- Facilitates efficient data transmission between parallel processors.
- Enables overlapping of computation and communication tasks.
- Optimizes traffic patterns for improved system performance.
Benefits:
- Enhanced system performance and efficiency.
- Improved data communication across devices.
- Increased parallel processing capabilities.
Commercial Applications: Potential commercial uses include:
- Supercomputing clusters.
- Cloud computing platforms.
- Data centers with distributed computing needs.
Prior Art: Readers can explore prior research on:
- Scheduling mechanisms in parallel processing systems.
- Communication optimization in distributed computing environments.
Frequently Updated Research: Stay informed about the latest advancements in:
- Parallel processing scheduling algorithms.
- Communication optimization techniques in distributed systems.
Questions about the Technology: 1. What are the key benefits of using a scheduling mechanism for reordering workgroups in a processing system? 2. How does this technology improve data transmission efficiency between parallel processors?
Original Abstract Submitted
A processing system includes a scheduling mechanism for producing data for fine-grained reordering of workgroups of a kernel to produce blocks of data, such as for communication across devices to enable overlapping of a producer computation with an all-reduce communication across the network. This scheduling mechanism enables a first parallel processor to schedule and execute a set of workgroups of a producer operation to generate data for transmission to a second parallel processor in a desired traffic pattern. At the same time, the second parallel processor schedules and executes a different set of workgroups of the producer operation to generate data for transmission in a desired traffic pattern to a third parallel processor or back to the first parallel processor.
