A user reported experiencing strange values in framebuffer memory usage for certain processes in a customer's Citrix environment with NVIDIA L4-4Q. They discovered that this issue can be caused by a broken Citrix XenDesktop graphics driver and was able to fix it by uninstalling and reinstalling the VDA. This issue has also been observed by Microsoft, who have a similar issue with GPU process memory counters reporting incorrect values. A support ticket has been opened for this issue and the ControlUp community has also seen similar cases.
Read the entire article here...
GPU Related Training & Support Archives
GPU training and support-related archives from inside the ControlUp Community on Slack.
Troubleshooting W365 CPC Integration and Visibility Issues
A group of ControlUp members discuss issues with the integration between W365 CPC and HDX. They note that EdgeDX does not detect brokering of a Cloud PC session, but the current configuration does show access to TURN and use of UDP. The group mentions that this may be due to the VDA configured on the device. There is also mention of potential confusion around which agent to use for the CPC Desktop, and a suggestion to pull a query to track the number of CPCs with a VDA installed and see what is reported. It is noted that there is interest from established enterprises to go this route.
Read the entire article here...
Read the entire article here...
Troubleshooting Strange GPU Usage in ControlUp’s Process and Machine Views
A user posts a question about their freshly installed environment and the strange GPU usage data they are getting from ControlUp. They provide details about their settings and ask for assistance in determining if the NVIDIA API is causing the issue. Other users suggest checking for shared memory usage and provide instructions on how to do so.
Read the entire article here...
Read the entire article here...
Troubleshooting vGPU Metrics with ControlUp
A discussion took place about vGPU metrics and how they differ depending on whether Windows or Nvidia API is used for data collection. One member asked for tips on making data collection via Nvidia API work and another member suggested checking for the installation of NVIDIA WMI and restarting the ControlUp agent after switching. ControlUp documentation on vGPU can be found at https://support.controlup.com/v1/docs/gpu-monitoring-with-controlup#collecting-data-from-nvidia-api-optional.
Read the entire article here...
Read the entire article here...
ControlUp Community Radio Episode #7 – Conversation with Ryan Ververs-Bijkerk & Eltjo van Gulik from Go-EUC.com
Troubleshooting High Peak User Input Delays in ControlUp
A user asked for recommendations to diagnose high Peak User Input Delays. Another user suggested zooming in on the time of the delay and checking memory usage. They also shared how to find information on what application is slow to respond to user input. The discussion then focused on how to identify the cause of a slow application, often due to resource constraints. The team provided examples, such as MMCexe processing event logs, and suggested that more resources may not always solve the issue.
Read the entire article here...
Read the entire article here...
Investigating Performance Issues in Citrix VDI with NVIDIA GRID GPU
A user reported strange performance differences between identical Citrix VDI machines with NVIDIA GRID GPU configured, all started from the same master. Though NVIDIA licensing and configuration was OK, some machines did not show a GPU in use. Other users suggested checking the ControlUp Agent delayed start setting and switching between native Microsoft APIs and NVIDIA metrics (https://support.controlup.com/docs/gpu-monitoring-with-controlup?highlight=gpu). They also recommended looking at known issues. The customer planned to investigate the issue further the next day.
Read the entire article here...
Read the entire article here...
Feature Requests to Expand ControlUp Reporting Capabilities
A user asked that ControlUp allow for Historic Data for GPU and Datastore latency and I/O historically in Reports via UserVoice suggestions, and urged others to upvote. ControlUp revealed they have already implemented Historic Data for GPU in some environments, and do not find granular Datastore data for latency and I/O useful since they don't look deeper than average/max over 5 minutes.
Read the entire article here...
Read the entire article here...
Does ControlUp Use WMI APIs?
The discussion provided an audit answer to the question of if the WMI APIs are used to manage virtual network resources. ControlUp usually uses them sparingly, such as for GPU data collection, but not for Monitor. A Script Action at https://www.controlup.com/script-library-posts/show-local-wmi-activity/ can be used to trace which process uses which WMI query.
Read the entire article here...
Read the entire article here...
Troubleshooting High GPU Memory Usage: How to Show GPU Memory Usage by Process?
A user noticed high GPU memory usage for VDI (80 to 95%) in the COP Console but when they selected processes, the GPU Utilization, Decoder Utilization, Encoder Utilization, and Frame Buffer Memory Utilization all showed 0%. They were wondering if there was a way to show the GPU memory usage by process. Another user suggested selecting System Sessiosn from the dropdown, but this was already done. It was suggested to check the GPU Utilization for more usage to narrow down the issue, but further investigation was needed.
Read the entire article here...
Read the entire article here...