Technology that adds this much value should be “capitalized”.

Now this may seem like a cheeky title but if you look into the value is of virtualizing the Graphical Processing Unit it makes perfect sense. Actually more like dollars and cents. Okay, enough with the word play for one blog post.

It used to be that graphic intensive applications were reserved for automotive designers, bridge engineers and other CAD/CAM power users. These days more and more diverse roles require the power of a GPU in modeling the data that they use. And one common way to address this was to provide these users with a local device (laptop) that had a dedicated GPU. Not only does this solution not scale well, it often leads to users storing sensitive data on that local device, close to the physical GPU.

A trip to the Doctor’s office

Take the example of my most recent visit to my doctor. He needed to go over the results of an MRI that was performed on my back. We sat down at his “desk”, he launched his MRI modeling app and opened my imaging file. He was able to twist and turn my spin images and explore it at angles that were honestly embarrassing. I’ve never looked at my tailbone from that perspective!

I will spare you the diagnosis. But it is what I didn’t see that was even more interesting, at least to me. When we sat down at his desk he didn’t have a physical computer, he had a thin client. When he launched that MRI modeling app he did it through Citrix Receiver. When he opened my MRI imaging file he did it from the data center drive. When the app quickly launched there was no delay from when he turned the image and the image responding. And when he selected areas, changed contrast, added color it responded with an accelerated graphics workstation-like experience. It was seamless.

Behind the scenes

Now let’s say, what he didn’t see and quite frankly as a medical doctor doesn’t care about, was what could have been happening in the data center. His IT department is running a blade server with XenServer and XenDesktop on top of an NVIDIA GRID vGPU. He launched his virtual desktop from receiver and accessed my medical record that was securely stored in the data center, next to the vGPU. His desktop took one of the available slices of the vGPU, which has 8 threads per. And because of the multi-threading capabilities of the vGPU in hosted virtual desktops, supported only on XenServer, his colleagues could do the same with their patients simultaneously.

All of this in a safe, secure, scalable, and centrally managed virtual environment. It also enables this office to provide workstation-like graphic performance from more economical thin clients. In fact, because of XenDesktop hosted virtual desktops and apps this appointment could have been performed anywhere, on any device. Imagine a doctor coming to your house with a tablet device to review your test results. It sounds so futuristic! {Insert sarcastic grin} Click here to Learn More.

This isn’t just limited to the medical field. Think of the possibilities in almost every industry: manufacturing, environmental, finance, engineering and so many more. If you had the ability to provide users with the power of a dedicated GPU, but with the economics and scale of hosted desktops, what would you do?  Get creative and write your answer in the comments.

How to get started:

  1. Check out this short video on CPU vs NVIDIA GRID K2 with XenDesktop
  2. Dig deeper into XenDesktop and NVIDIA GRID