Applications By GPU - AVD Test

GPU-accelerated computing is the use of a graphics processing unit (GPU) together with a CPU to accelerate scientific, analytics, engineering, consumer, and enterprise applications. GPU-accelerated computing enhances application performance by offloading compute-intensive portions of the application to the GPU, while the remainder of the code still runs on the CPU.

In GPU-enabled AVD environments, if users accessing graphic virtual applications complain of slowness , administrators must be able to instantly figure out what is causing the slowness – is it because adequate GPU resources are not available to the host? Or is it because of excessive utilization of GPU memory and processing resources by a few virtual applications on the host?  Accurate answers to these questions can help administrators determine whether/not:

  • The host is sized with sufficient GPU resources;
  • The GPUs are configured with enough graphics memory;

Measures to right-size the host and fine-tune its GPU configuration can be initiated based on the results of this analysis. This is exactly what the Applications By GPU - AVD test helps administrators achieve! 

To help with better utilization of resources, you can track the GPU usage rates of your instances. When you know the GPU usage rates, you can then perform tasks such as setting up managed instance groups that can be used to autoscale resources based on needs.

Target of the test : An Azure Virtual Desktop

Agent deploying the test : An internal agent.

Outputs of the test : One set of results for each application on the chosen Session Host / Azure Virtual Desktop that is using GPU

Configurable parameters for the test
Parameters Description

Test Period

How often should the test be executed.

Host

The host for which the test is to be configured.

Port

The default port is NULL.

NVIDIA Home

This test uses NVIDIA WMI (NVWMI) to pull metrics on GPU usage by applications. For this test to run therefore, you need to install NVWMI on the target Azure Virtual Desktop. If the NVWMI is installed in its default directory, then this test will automatically discover the location of NVWMI and use it for pulling the desired metrics. This is why, the NVIDIA Home parameter is set to none by default. However, if you have installed NVWMI in a different directory, then you have to explicitly configure the path to that installation directory against the NVIDIA Home parameter.

Report By Domain Name

By default, this flag is set to Yes. This implies that by default, the detailed diagnosis of this test will display the domainname\username of each user who accessed an application on the AVD. This way, administrators will be able to quickly determine which user logged into the AVD from which domain. If you want the detailed diagnosis to display only the user name of these users, set this flag to No.

DD Frequency

Refers to the frequency with which detailed diagnosis measures are to be generated for this test. The default is 1:1. This indicates that, by default, detailed measures will be generated every time this test runs, and also every time the test detects a problem. You can modify this frequency, if you so desire. Also, if you intend to disable the detailed diagnosis capability for this test, you can do so by specifying none against DD Frequency

Detailed Diagnosis

To make diagnosis more efficient and accurate, the eG Enterprise embeds an optional detailed diagnostic capability. With this capability, the eG agents can be configured to run detailed, more elaborate tests as and when specific problems are detected. To enable the detailed diagnosis capability of this test for a particular server, choose the On option. To disable the capability, click on the Off option.

The option to selectively enable/disable the detailed diagnosis capability will be available only if the following conditions are fulfilled:

  • The eG manager license should allow the detailed diagnosis capability
  • Both the normal and abnormal frequencies configured for the detailed diagnosis measures should not be 0.
Measurements made by the test
Measurement Description Measurement Unit Interpretation

GPU instances currently running

Indicates the number of instances of this graphic application that are currently running.

Number

 

GPU compute usage

Indicates the percentage of time for which this application was utilizing GPU.

Percent

A value close to 100% indicates that the GPU is busy processing graphic requests from this application almost all the time.

In such a case, you may want to consider allocating more GPU resources to that AVD.

Encoder usage

Indicates the amount of GPU that this application is utilizing for the encoding process.

Percent

A value close to 100 is a cause of concern. By closely analyzing these measures, administrators can easily be alerted to situations where graphics processing is a bottleneck.

 

Decoder usage

Indicates the amount of GPU that this application is utilizing for the decoding process.

Percent

Memory compute usage

Indicates the percentage of time during which memory on the GPU was read from/written to by this application.

Percent

A value close to 100% is a cause for concern as it indicates that the application is almost always using the graphics memory on a GPU.

In such a case, you may want to consider allocating more graphics memory to that AVD.

Memory used

Indicates the amount of graphics memory used by this application.

MiB

Compare the value of this measure across applications to know which application is hogging GPU memory.