Summary

This guide serves as a starting point for right sizing your enterprise deployments for knowledge workers. When sizing a knowledge worker VDI deployment with NVIDIA vGPU, NVIDIA recommends conducting a POC and thoroughly analyzing resource utilization using objective measurements.

If a customer is considering using L40/L40S, L4, or the RTX PRO 6000 Blackwell Server Edition for specific workloads, it’s essential to note that the framebuffer footprint per GPU will be higher compared to A16s. The RTX PRO 6000 Blackwell Server Edition offers up to 96 GB of framebuffer and supports higher user density with enhanced performance, but requires careful sizing. Therefore, before implementing these alternatives, it is advisable to conduct a POC to gain a better understanding of performance, user density, and cost-effectiveness in your environment.

The best effort scheduler option is recommended for enterprise deployments, and user density will be dependent on the hardware configuration, user profile, workloads and display resolution.

To see how you can virtualize digital knowledge worker workloads using NVIDIA vPC software, try it for free.