In the latest GeekOut 365 video showcase, Patrick van den Born, a senior consultant at PQR, shows off how he determined GPU sizing for a client’s VDI project.
Patrick had an aerospace supplier client looking to move GPU-intensive workloads from their current fat clients to Citrix XenDesktop with NVIDIA Tesla M60s, without sacrificing the current user experience. So, he needed to figure out the appropriate GPU size the client would need, which isn’t always easy.
In the video, Patrick explains how he used Splunk and uberAgent to collect data on the client’s current resource utilization. Given the client used two different GPU vendors (Intel and NVIDIA), he normalized the data and narrowed his focus to GFLOPS. From there, he took the data collected and compared current GFLOPS and framebuffer.
With the info that Splunk and uberAgent provided, Patrick was able to determine the ideal VDI and GPU setup for his clients. Watch the TeamRGE video to learn more about his process and use his takeaways in determining your or a client’s ideal GPU needs.
Next week: It’s August, already. Also, we’ll be highlighting a public cloud smackdown!