Method, device and computer readable medium for scheduling dedicated processing resource

Patent metadata

Field	Value
Publication number	US-11061731-B2
Application number	US-201916389166-A
Country	US
Kind code	B2
Filing date	Apr 19, 2019
Priority date	Apr 20, 2018
Publication date	Jul 13, 2021
Grant date	Jul 13, 2021

Abstract

A method of scheduling a dedicated processing resource includes: obtaining source code of an application to be compiled; extracting, during compiling of the source code, metadata associated with the application, the metadata indicating an amount of the dedicated processing resource required by the application; and obtaining, based on the metadata, the dedicated processing resource allocated to the application. In this manner, performance of the dedicated processing resource scheduling system and resource utilization is improved.

First claim

What is claimed is:

1. A method of scheduling a dedicated processing resource, comprising:
obtaining source code of an application to be compiled;
extracting, during compiling of the source code, metadata through an extraction function embedded in a compiler for compiling the source code, the metadata being associated with the application, and the metadata indicating an amount of the dedicated processing resource required by the application; and
obtaining, based on the metadata, the dedicated processing resource allocated to the application.

2. The method of claim 1 , wherein obtaining the dedicated processing resource allocated to the application comprises:
analyzing the metadata to predict the dedicated processing resource required by the application;
requesting the dedicated processing resource from a remote controller; and
receiving a dedicated processing resource notification from the remote controller, the dedicated processing resource notification indicating the dedicated processing resource allocated by the remote controller to the application.

3. The method of claim 1 , wherein obtaining the dedicated processing resource allocated to the application comprises:
sending the metadata to a remote database;
requesting the dedicated processing resource from a remote controller to enable the controller to access the remote database and analyze the metadata; and
receiving a dedicated processing resource notification from the remote controller, the dedicated processing resource notification indicating the dedicated processing resource allocated by the remote controller to the application.

4. The method of claim 1 , wherein the application comprises a deep learning application, and wherein the metadata comprises at least one of:
a type of at least one layer in a model of the deep learning application;
the number of layers in the model of the deep learning application; and
a format of data input to the deep learning application.

5. The method of claim 1 , wherein the dedicated processing resource required by the application is a graphical processing unit (GPU), and wherein the metadata comprises at least one of:
a number of kernels of the GPU required by the application;
an amount of a computing resource of the GPU required by the application; and
an amount of a memory resource of the GPU required by the application.

6. The method of claim 5 , wherein the application obtains the required GPU from a GPU resource pool via a network connection.

7. The method of claim 1 , wherein extracting the metadata associated with the application comprises:
obtaining a journal generated during compiling of the source code; and
extracting, based on the journal, the metadata associated with the application.

8. A device for scheduling a dedicated processing resource, comprising:
at least one processing unit; and
at least one memory coupled to the at least one processing unit and storing instructions executed by the at least one processing unit, the instructions, when executed by the at least one processing unit, causing the device to execute acts, the acts comprising:
obtaining source code of an application to be compiled;
extracting, during compiling of the source code, metadata through an extraction function embedded in a compiler for compiling the source code, the metadata being associated with the application, and the metadata indicating an amount of the dedicated processing resource required by the application; and
obtaining, based on the metadata, the dedicated processing resource allocated to the application.

9. The device of claim 8 , wherein obtaining the dedicated processing resource allocated to the application comprises:
analyzing the metadata to predict the dedicated processing resource required by the application;
requesting the dedicated processing resource from a remote controller; and
receiving a dedicated processing resource notification from the remote controller, the dedicated processing resource notification indicating the dedicated processing resource allocated by the remote controller to the application.

10. The device of claim 8 , wherein obtaining the dedicated processing resource allocated to the application comprises:
sending the metadata to a remote database;
requesting the dedicated processing resource from a remote controller to enable the controller to access the remote database and analyze the metadata; and
receiving a dedicated processing resource notification from the remote controller, the dedicated processing resource notification indicating the dedicated processing resource allocated by the remote controller to the application.

11. The device of claim 8 , wherein the application comprises a deep learning application, and wherein the metadata comprise at least one of:
a type of one of layers in a model of the deep learning application;
the number of the layers in the model of the deep learning application; and
a format of data input to the deep learning application.

12. The device of claim 8 , wherein the dedicated processing resource required by the application is a graphical processing unit (GPU), and wherein the metadata comprises at least one of:
a number of kernels of the GPU required by the application;
an amount of a computing resource of the GPU required by the application; and
an amount of a memory resource of the GPU required by the application.

13. The device of claim 12 , wherein the application obtains the required GPU from a GPU resource pool via a network connection.

14. The device of claim 8 , wherein extracting the metadata associated with the application comprises:
obtaining a journal generated during compiling of the source code; and
extracting, based on the journal, the metadata associated with the application.

15. A computer readable program product, which stores machine executable instructions thereon, the machine executable instructions, when executed by at least one processor, causing the at least one processor to implement a method of scheduling a dedicated processing resource, comprising:
obtaining source code of an application to be compiled;
extracting, during compiling of the source code, metadata through an extraction function embedded in a compiler for compiling the source code, the metadata being associated with the application, and the metadata indicating an amount of the dedicated processing resource required by the application; and
obtaining, based on the metadata, the dedicated processing resource allocated to the application.

16. The computer readable program product of claim 15 , wherein obtaining the dedicated processing resource allocated to the application comprises:
analyzing the metadata to predict the dedicated processing resource required by the application;
requesting the dedicated processing resource from a remote controller; and
receiving a dedicated processing resource notification from the remote controller, the dedicated processing resource notification indicating the dedicated processing resource allocated by the remote controller to the application.

17. The computer readable program product of claim 15 , wherein obtaining the dedicated processing resource allocated to the application comprises:
sending the metadata to a remote database;
requesting the dedicated processing resource from a remote controller to enable the controller to access the remote database and analyze the metadata; and
receiving a dedicated processing resource notification from the remote controller, the dedicated processing resource notification indicating the dedicat

Assignees

• Emc Ip Holding Co Llc

Inventors

• Zhao Junping
• Wang Kun
• Peng Layne Lin
• Chen Fei

Classifications

• G06F8/43 Primary

  Checking; Contextual analysis · CPC title

• G06F9/505 Primary

  considering the load · CPC title

• G06F9/5027

  the resource being a machine, e.g. CPUs, Servers, Terminals · CPC title

• G06F8/41

  Compilation · CPC title