Revving Up Graphics and Virtual Machines
Written on July 8, 2009 by Katherine W. Prawl
Working on a forthcoming review of vReveal video enhancment software (for Windows®) has led me to learn more about both Graphical Processing Units (GPUs) and virtual machine (VM) applications, not to mention brushing up on various recent operating systems. When I realized the opensource VM app I was using (Sun’s VirtualBox) wasn’t taking advantage of my iMac’s video card (an Nvidia GeForce MT 130), I dug into the manual to see if there was a way to improve its performance. There actually is a way to make the VMs work better, called VirtualBox Guest Additions, which includes a lot of extra drivers and utilities, but sadly that doesn’t include drivers for any Nvidia® CUDA™enhanced GPUs.
However, while it is not possible to get the hardware accelleration normally available from the GPU, it is still possible to use vReveal to clean up shaky, low-contrast videos. It just takes longer since it will be processed by the CPU instead of the GPU. Much longer — in some cases it can be up to 20Xs slower on the CPU. But if you have the time, it can be done with just one click of a button.
While investigating CUDA, I was bemused to see a whitepaper in the Nvidia CUDA Zone called “Fast N-body Simulation with CUDA”. If you’ve been following this blog for awhile, you may recall an item I wrote about a year ago, called Virtual Worlds Galore, that included a video made by the guys at Genkii in cooperation with the Meta Institute for Computational Astrophysics, or MICA, on this subject. Apparently n-body simulations are all the rage now that desktop computers are capable of displaying them, thanks to virtualization and graphics acceleration architectures like the Nvidia CUDA. They’re a bit over my head, since I’m neither an astrophysicist nor a 3D visualization expert, but its fascinating to follow what’s being done by people who are. And it is more than a little interesting to see how so many of the topics I cover are connected one way or another in unexpected ways.
If you enjoyed this post Subscribe to our feed
