Hypervisor scheduling of workloads on heterogeneous processing cores based on workload hints

Patent metadata

Field	Value
Publication number	US-2025068450-A1
Application number	US-202318456161-A
Country	US
Kind code	A1
Filing date	Aug 25, 2023
Priority date	Aug 25, 2023
Publication date	Feb 27, 2025
Grant date	—

Abstract

In one set of embodiments, hint metadata for a workload may be provided to a hypervisor, where the hint metadata includes information pertaining to the computational nature, preferences, and/or requirements of the workload (or some portion thereof). This hint metadata may be provided statically prior to VM power-on and/or dynamically during VM runtime. The hypervisor may then schedule the workload on one or more of a plurality of heterogenous processing cores based on the hint metadata.

First claim

What is claimed is:

1 . A method comprising:
receiving, by a hypervisor of a host system, hint metadata for a workload running on the hypervisor, wherein the host system includes a plurality of heterogeneous processing cores, and wherein the hint metadata indicates a computational nature, need, and/or preference of at least a portion of the workload; and scheduling, by the hypervisor, the workload on one or more of the plurality of heterogeneous processing cores based on the hint metadata.

2 . The method of claim 1 wherein the hint metadata was statically assigned to the workload prior to initiation of the workload.

3 . The method of claim 1 wherein the hint metadata is dynamically determined while the workload is running.

4 . The method of claim 3 wherein the hint metadata is dynamically determined by an agent running within the workload that:
detects a launch of a workload process;
matches the workload process to an entry in a process list;
retrieves a tag for the workload process from the matched entry, wherein the tag indicates a computational nature, need, and/or preference of the workload process;
determines a computational unit of the workload to which the workload process is mapped; and
sends the tag and an identifier of the computational unit to the hypervisor.

5 . The method of claim 4 wherein the hypervisor schedules the computational unit for execution on one of the plurality of heterogeneous processing cores based on the tag.

6 . The method of claim 1 wherein the plurality of heterogeneous processing cores includes performance cores and efficiency cores, wherein the hint metadata indicates a priority level of the workload, and wherein the workload is scheduled on one or more of the performance cores and/or one or more of the efficiency cores in accordance with the priority level.

7 . The method of claim 1 wherein the plurality of heterogeneous processing cores includes specialized processing cores designed to perform a particular task, wherein the hint metadata indicates that the workload is directed to the particular task, and wherein the workload is scheduled on one or more of the specialized processing cores.

8 . The method of claim 1 wherein the scheduling fails due to a lack of availability of appropriate processing cores on the host system, and wherein in response to this failure a virtual infrastructure management server migrates the VM to another host system that can accommodate the workload in accordance with the hint metadata.

9 . A non-transitory computer readable storage medium having stored thereon program code executable by a hypervisor of a computer system, the program code causing the hypervisor to execute a method comprising:
receiving hint metadata for a workload running on the hypervisor, wherein the computer system includes a plurality of heterogeneous processing cores, and wherein the hint metadata indicates a computational nature, need, and/or preference of at least a portion of the workload; and scheduling the workload on one or more of the plurality of heterogeneous processing cores based on the hint metadata.

10 . The non-transitory computer readable storage medium of claim 9 wherein the hint metadata was statically assigned to the workload prior to initiation of the workload.

11 . The non-transitory computer readable storage medium of claim 9 wherein the hint metadata is dynamically determined while the workload is running.

12 . The non-transitory computer readable storage medium of claim 11 wherein the hint metadata is dynamically determined by an agent running within the workload that:
detects a launch of a workload process;
matches the workload process to an entry in a process list;
retrieves a tag for the workload process from the matched entry, wherein the tag indicates a computational nature, need, and/or preference of the workload process;
determines a computational unit of the workload to which the workload process is mapped; and
sends the tag and an identifier of the computational unit to the hypervisor.

13 . The non-transitory computer readable storage medium of claim 12 wherein the hypervisor schedules the computational unit for execution on one of the plurality of heterogeneous processing cores based on the tag.

14 . The non-transitory computer readable storage medium of claim 9 wherein the plurality of heterogeneous processing cores includes performance cores and efficiency cores, wherein the hint metadata indicates a priority level of the workload, and wherein the workload is scheduled on one or more of the performance cores and/or one or more of the efficiency cores in accordance with the priority level.

15 . The non-transitory computer readable storage medium of claim 9 wherein the plurality of heterogeneous processing cores includes specialized processing cores designed to perform a particular task, wherein the hint metadata indicates that the workload is directed to the particular task, and wherein the workload is scheduled on one or more of the specialized processing cores.

16 . The non-transitory computer readable storage medium of claim 9 wherein the scheduling fails due to a lack of availability of appropriate processing cores on the host system, and wherein in response to this failure a virtual infrastructure management server migrates the VM to another host system that can accommodate the workload in accordance with the hint metadata.

17 . A computer system comprising:
a plurality of heterogeneous processing cores; a hypervisor; and a non-transitory computer readable medium having stored thereon program code that, when executed, causes the hypervisor to:
receive hint metadata for a workload running on the hypervisor, wherein the hint metadata indicates a computational nature, need, and/or preference of at least a portion of the workload; and
schedule the workload on one or more of the plurality of heterogeneous processing cores based on the hint metadata.

18 . The computer system of claim 17 wherein the hint metadata was statically assigned to the workload prior to initiation of the workload.

19 . The computer system of claim 17 wherein the hint metadata is dynamically determined while the workload is running.

20 . The computer system of claim 18 wherein the hint metadata is dynamically determined by an agent running within the workload that:
detects a launch of a workload process;
matches the workload process to an entry in a process list;
retrieves a tag for the workload process from the matched entry, wherein the tag indicates a computational nature, need, and/or preference of the workload process;
determines a computational unit of the workload to which the workload process is mapped; and
sends the tag and an identifier of the computational unit to the hypervisor.

21 . The computer system of claim 19 wherein the hypervisor schedules the computational unit for execution on one of the plurality of heterogeneous processing cores based on the tag.

22 . The computer system of claim 17 wherein the plurality of heterogeneous processing cores includes performance cores and efficiency cores, wherein the hint metadata indicates a priority level of the workload, and wherein the workload is scheduled on one or more of the performance cores and/or one or more of the efficiency cores in accordance with the priority level.

23 . The computer system of claim 17 wherein the plurality of heterogeneous processing cores

Assignees

• VMware LLC

Inventors

• Beveridge Daniel James
• Lyons Corey Douglas

Classifications

• G06F9/4881

  Scheduling strategies for dispatcher, e.g. round robin, multi-level priority queues · CPC title

• G06F2009/4557

  Distribution of virtual machine instances; Migration and load balancing · CPC title

• G06F9/45558 Primary

  Hypervisor-specific management and integration aspects · CPC title