Method and system for recovering data associated with artificial intelligence calculation

1. A method for recovering data for artificial intelligence calculation, the method being performed by one or more processors and comprising:

extracting, from a command queue, a descriptor associated with a target job, that is a job to be executed, of a plurality of jobs;

executing at least one command associated with the extracted descriptor to execute artificial intelligence calculation associated with the target job;

resetting the command queue if an error occurs while executing the artificial intelligence calculation;

determining at least one descriptor to be recovered, based on tracking data associated with at least one job; and

recovering the determined at least one descriptor to the reset command queue.

2. The method according to claim 1 , wherein the determining at least one descriptor to be recovered comprises:

identifying a descriptor associated with the error; and

determining at least one descriptor subsequent to the identified descriptor to be a descriptor to be recovered.

3. The method according to claim 1 , wherein the determining at least one descriptor to be recovered comprises:

identifying a descriptor associated with the error; and

determining the identified descriptor and at least one descriptor subsequent to the identified descriptor to be descriptors to be recovered.

4. The method according to claim 1 , wherein the determining at least one descriptor to be recovered comprises:

identifying a descriptor associated with the error;

identifying a number of errors in the artificial intelligence calculation associated with the identified descriptor; and

if the number of identified errors exceeds a threshold, determining at least one descriptor subsequent to the identified descriptor to be a descriptor to be recovered.

5. The method according to claim 1 , wherein the recovering the determined at least one descriptor to the reset command queue comprises:

obtaining, from the tracking data, at least one descriptor stored in the command queue before reset; and

recovering the command queue based on the obtained at least one descriptor.

6. The method according to claim 1 , wherein the recovering the determined at least one descriptor to the reset command queue comprises:

identifying, based on the tracking data, at least one command buffer address associated with the command queue before reset; and

recovering the command queue based on a command stored in the identified at least one command buffer address.

7. The method according to claim 1 , further comprising, prior to the resetting the command queue:

identifying context associated with a job having the error; and

controlling the command queue such that an additional descriptor associated with the identified context is not stored in the command queue.

8. The method according to claim 7 , wherein the context is associated with at least one of a specific user or a specific port.

9. The method according to claim 1 , further comprising, prior to the resetting the command queue:

identifying context associated with a job having the error; and

if an additional job occurs that is not associated with the identified context, storing a descriptor associated with the additional job in the command queue.

10. The method according to claim 1 , wherein, at least one job data associated with a specific context is stored in a job queue,

a descriptor stored in the command queue is associated with specific job data stored in the job queue, and

a command associated with a specific descriptor stored in the command queue is stored in a command buffer.

11. The method according to claim 1 , wherein the executed at least one command is associated with inference performed in a neural processing device.

12. A non-transitory computer-readable recording medium storing instructions that, when executed by one or more processors, cause performance of the method according to claim 1 .

13. A computing system comprising:

a command queue storing at least one descriptor; and

one or more processors configured to manage the command queue,

wherein the one or more processors are further configured to:

extract, from the command queue, a descriptor associated with a target job, that is a job to be executed, of a plurality of jobs;

execute at least one command associated with the extracted descriptor to execute artificial intelligence calculation associated with the target job;

reset the command queue if an error occurs while executing the artificial intelligence calculation;

determine at least one descriptor to be recovered, based on tracking data associated with at least one job; and

recover the determined at least one descriptor to the reset command queue.

14. The computing system according to claim 13 , wherein the determining at least one descriptor to be recovered comprises:

identifying a descriptor associated with the error; and

determining at least one descriptor subsequent to the identified descriptor to be a descriptor to be recovered.

15. The computing system according to claim 13 , wherein the determining at least one descriptor to be recovered comprises:

identifying a descriptor associated with the error; and

determining the identified descriptor and at least one descriptor subsequent to the identified descriptor to be descriptors to be recovered.

16. The computing system according to claim 13 , wherein the determining at least one descriptor to be recovered comprises:

identifying a descriptor associated with the error; and

identifying a number of errors in the artificial intelligence calculation associated with the identified descriptor; and

if the number of identified errors exceeds a threshold, determining at least one descriptor subsequent to the identified descriptor to be a descriptor to be recovered.

17. The computing system according to claim 13 , wherein the recovering the determined at least one descriptor to the reset command queue comprises:

obtaining, from the tracking data, at least one descriptor stored in the command queue before reset; and

recovering the command queue based on the obtained at least one descriptor.

18. The computing system according to claim 13 , wherein the recovering the determined at least one descriptor to the reset command queue comprises:

identifying, based on the tracking data, at least one command buffer address associated with the command queue before reset; and

recovering the command queue based on a command stored in the identified at least one command buffer address.

19. The computing system according to claim 13 , wherein the one or more processors are further configured to:

before resetting the command queue, identify a context associated with a job having the error, and

control the command queue such that an additional descriptor associated with the identified context is not stored in the command queue.

20. The computing system according to claim 13 , wherein the one or more processors are further configured to:

before resetting the command queue, identify a context associated with a job having the error, and

if an additional job occurs that is not associated with the identified context, store a descriptor associated with the additional job in the command queue.