Method and system for processing data using a processing pipeline and processing units
A method for processing data includes receiving an offload request by a first virtual machine (VM), issuing, in response to the offload request and based on a processing pipeline, a processing request to a processing unit, and servicing, by the processing unit, the processing request to obtain a result.
1. A method for processing data, the method comprising:
receiving an offload request by a first virtual machine (VM), wherein the offload request is obtained from an offload component executing in a host operating system;
determining a processing pipeline from a plurality of processing pipelines;
issuing, in response to the offload request and based on the processing pipeline, a processing request to a processing unit, wherein the processing request comprises reference to data;
servicing, by the processing unit, the processing request,
wherein servicing the processing request comprises:
obtaining the data from shared memory using the reference to the data;
generating a result using the data and using the processing pipeline; and
storing the result in the shared memory;
obtaining, by the host operating system, the result from the shared memory;
combining the result with the data to obtain an updated result; and
providing the updated result to an external computing device,
wherein the first VM, the host operating system, and the processing unit are executing on a computing device.
2. The method of claim 1 , further comprising:
receiving a second offload request by a second VM executing on the computing device;
issuing, in response to the second offload request and based on a second processing pipeline, a second processing request to the processing unit; and
servicing, by the processing unit, the second processing request to obtain a second result.
3. The method of claim 2 , further comprising:
prior to issuing the second offload request, obtaining the second processing pipeline from a host operating system by the second VM, wherein the second VM is executing on the host operating system.
4. The method of claim 1 , further comprising:
issuing, in response to the offload request and based on the processing pipeline, a second processing request to the processing unit; and
servicing, by the processing unit, the second processing request to obtain a second result, wherein the second result is based on the result.
5. The method of claim 4 , wherein the processing pipeline specifies at least a first processing task and a second processing task, wherein the processing request is associated with the first processing task and the second processing request is associated with the second processing task.
6. The method of claim 1 , wherein the offload request is issued by an offload component executing in a host operating system, wherein the first VM is executing on the host operating system.
7. The method of claim 6 , wherein the offload request is issued in response to a client request issued by a third VM executing on the host operating system.
8. The method of claim 1 , wherein the processing unit is a graphics processing unit (GPU).
9. The method of claim 1 , wherein the processing pipeline specifies a plurality of processing tasks.
10. The method of claim 9 , wherein at least one processing task of the plurality of tasks is compression or a data integrity calculation.
11. A non-transitory computer readable medium comprising computer readable program code, which when executed by a computer processor enables the computer processor to perform a method for processing data, the method comprising:
receiving an offload request by a first virtual machine (VM), wherein the offload request is obtained from an offload component executing in a host operating system;
determining a processing pipeline from a plurality of processing pipelines;
issuing, in response to the offload request and based on the processing pipeline, a processing request to a processing unit, wherein the processing request comprises reference to data;
servicing, by the processing unit, the processing request,
wherein servicing the processing request comprises:
obtaining the data from shared memory using the reference to the data;
generating a result using the data and using the processing pipeline; and
storing the result in the shared memory;
obtaining, by the host operating system, the result from the shared memory;
combining the result with the data to obtain an updated result; and
providing the updated result to an external computing device,
wherein the first VM, the host operating system, and the processing unit are executing on a computing device.
12. The non-transitory computer readable medium of claim 11 , wherein the method further comprises:
issuing, in response to the offload request and based on the processing pipeline, a second processing request to the processing unit,
wherein the processing unit services the second processing request to obtain a second result, wherein the second result is based on the result.
13. The non-transitory computer readable medium of 12 , wherein the processing pipeline specifies at least a first processing task and a second processing task, wherein the processing request is associated with the first processing task and the second processing request is associated with the second processing task.
14. A system, comprising:
a processor;
a plurality of processing units;
shared memory;
a host operating system executing on the processor;
a plurality of virtual machines (VMs) executing on the host operating system;
wherein a first VM of the plurality of VMs is programmed to:
receive an offload request from the host operating system;
determining a processing pipeline of a plurality of processing pipelines;
issue, in response to the offload request and based on the processing pipeline, a processing request to a processing unit of the plurality of processing units, wherein the processing request comprises referenc to data; and
wherein the processing unit is configured to service the processing request to obtain a result and store the result in the shared memory,
wherein servicing the processing request comprises:
obtaining the data from the shared memory using the reference to the data;
generating the result using the data; and
storing the result in the shared memory;
wherein the host operating system is programmed to:
obtain the result from the shared memory;
combining the result with the data to obtain an updated result; and
providing the updated result to an external computing device.
15. The system of claim 14 , wherein the processing unit is a graphics processing unit (GPU).
16. The system of claim 15 , wherein a processing task of the plurality of processing tasks is one of: a compression or a data integrity calculation.
17. The system of claim 14 , wherein the processing pipeline specifies a plurality of processing tasks.