17878677. SCHEDULING OPTIMIZATION IN SEQUENCE SPACE simplified abstract (QUALCOMM Incorporated)

SCHEDULING OPTIMIZATION IN SEQUENCE SPACE

Organization Name

QUALCOMM Incorporated

Inventor(s)

Yang Yang of San Diego CA (US)

Mukul Gagrani of San Diego CA (US)

Wonseok Jeon of San Diego CA (US)

Edward Teague of San Diego CA (US)

Weiliang Zeng of San Diego CA (US)

Piero Zappi of La Jolla CA (US)

Corrado Rainone of Haarlem (NL)

Christopher Lott of San Diego CA (US)

SCHEDULING OPTIMIZATION IN SEQUENCE SPACE - A simplified explanation of the abstract

This abstract first appeared for US patent application 17878677 titled 'SCHEDULING OPTIMIZATION IN SEQUENCE SPACE

Simplified Explanation

The abstract of this patent application describes a method for generating a schedule to execute operations of a compute graph. The method involves receiving a graph with multiple nodes connected by edges, where each node represents an operation to be executed. The method determines a set of sequences for executing the nodes based on precedence constraints. It then selects one or more sequences from the set based on a memory constraint associated with the device executing the nodes. Finally, a schedule for executing the nodes on the device is generated based on the selected sequences.

• The patent application describes a method for scheduling the execution of operations in a compute graph.
• The method takes into account precedence constraints and memory constraints to determine the order of execution.
• The method generates a schedule based on selected sequences that satisfy the memory constraint.
• The innovation allows for efficient execution of operations in a compute graph, optimizing resource utilization.

Potential applications of this technology:

• This technology can be applied in various fields where compute graphs are used, such as machine learning, data processing, and scientific simulations.
• It can be used in distributed computing systems to optimize the execution of operations across multiple devices.
• This technology can improve the performance and efficiency of parallel computing systems.

Problems solved by this technology:

• The method solves the problem of efficiently scheduling the execution of operations in a compute graph.
• It addresses the challenge of considering both precedence and memory constraints to optimize the execution order.
• This technology solves the problem of resource underutilization and inefficient execution in compute graphs.

Benefits of this technology:

• The method allows for improved resource utilization by optimizing the execution order of operations.
• It can lead to faster execution times and improved performance in compute-intensive tasks.
• This technology enables efficient utilization of memory resources, reducing memory bottlenecks and improving overall system efficiency.

Original Abstract Submitted

A processor-implemented method for generating a schedule for executing operations of a compute graph includes receiving a graph including multiples nodes connected by edges. Each of the multiple nodes represents an operation to be executed. A set of sequences for executing the nodes is determined based on one or more precedence constraints. One or more sequences are selected from the set of sequences based on a memory constraint associated with a device for executing the nodes. A schedule for executing the nodes on the device is generated based on the selected one or more sequences.