Home > Workload Solutions > Virtual Desktop Infrastructure > White Papers > Technical White Paper—VDI Graphics Acceleration on a Dell MX7000 Modular Chassis > Load generation
The default workload for the NVIDIA nVector suite is the Knowledge Worker profile. However, we used the more demanding SPECviewperf 13 professional graphics workload with the nVector tool for this performance testing. The SPECviewperf 13 benchmark measures the 3D graphics performance of systems running under the OpenGL and Direct X application programming interfaces. Because SPECviewperf 13 is a standard benchmark for measuring graphics performance for professional applications, including 3ds Max and Maya, it was the most appropriate choice for this testing effort.
The SPECviewperf 13 workload uses a series of viewsets taken from independent software vendor (ISV) applications to characterize the graphics performance of a physical or virtual workstation. It uses GPU-enabled launchers or endpoint VMs to connect a specified number of users to the available virtual desktops within the environment. After the simulated user is connected, a login script starts a test workload script on the user environment that is configured by the login script. Each launcher or endpoint system can launch connections to a single “target” machine (such as a VDI desktop). Mapping takes place from the endpoint to the desktop used by a specific user. The launchers and desktops are deployed and monitored using the nVector management VM where the framework runs. The data collected during the test is analyzed afterwards.
The SPECviewperf 13 benchmark’s workloads, called viewsets, represent graphics content and behavior typical of actual applications. We used the Autodesk 3ds Max 2016 (3dsmax-06) and Autodesk Maya 2017 (Maya-05) viewsets that are available within the benchmark. For more information about the SPECviewperf 13 viewsets, see the SPEC documentation.
We used the following login and boot scheduling paradigm: