May 26, 2014
I was reading some SIB’s (self-important blogger’s) bloviation about what he saw at the Consumer Electronics Show this winter. The number and quality of the 4K TVs on display blew him away. 4K is techie hipster lingo for ultra high-definition (3840 x 2160 pixels). To put that in perspective: Your high-def monitor or power wall at work is probably in the 1080 pixel range like the HDTV at home.
Now imagine a 33.2 megapixel 8K IUHD (insanely ultra high definition) power wall displaying 7680 x 4320 pixels. OK, I made up IUHD, but not the 8K spec. It’s coming. Soon. Are you ready? Today’s Check it Out link takes you to just such an outfit with a brand-spanking new workstation that can handle 8K even though 4K is not widely deployed yet.
But, as an aside, you can credit NVIDIA for a lot of this glorious graphics insanity. They figured out how to couple GPUs (graphics processing units) with CPUs so that the graphics- and compute-intensive parts of your applications run on the GPU while your CPU handles the rest. The upshot is that GPU-accelerated computing runs rings around the old way of doing things, especially when you run engineering and design applications coded for it. There are 200- something applications optimized for GPU-accelerated computing. In engineering design, analysis and rendering, that growing list includes Abaqus, ANSYS Fluent and Mechanical, CST Microwave Studio, MATLAB, and 3ds Max.
Still, even if NVIDIA figured out the GPU-accelerated computing, it takes a savvy manufacturer working in cahoots with them to build a workstation that can handle that potential properly. Rave Computer, the company behind today’s Check it Out subject, is just such an outfit. It has a brand new workstation call the Rave RT-1250 that leverages NVIDIA GPU technology in a big way.
Well, maybe it’s better to describe Rave Computer’s RT-1250 as a desktop-sized supercomputer. Here are a few quick reasons why it’s appropriate to call the RT-1250 a supercomputer.
The Rave RT-1250 supports dual Intel Xeon E5-2600 v2 series processors – those are the CPUs that you can build a data center around. It supports up to four double-width graphics cards or coprocessors, including professional-level NVIDIA Quadro graphics cards with one or more NVIDIA Tesla K40 GPUs. The Tesla K40 GPU is engineered for HPC (high-performance computing) and Big Data jobs, according to NVIDIA. They offer 1.4 TFLOPS performance, 12GB memory and 288 GB/s throughput.
So what does all this mean for you? First off, it means that the desktop-sized Rave RT-1250 can supply you with the power to process your huge 3D modeling, CAD, CFD (computational fluid dynamics), numerical analytics, simulation and visualization jobs so efficiently that it’ll blow you away. For that matter, it can bring the same sort of efficiency to coping with those massive datasets the financial people, chemists and digital media artists are into as well.
Secondly, Rave Computer has made its mark over the past couple of decades in part by custom-engineering solutions optimized for industrial computing, military grade, robotics, simulation, visualization and other tightly specific customer requirements. For example, the recently launched RT-1250 made its debut in a form optimized to power an enterprise 8K by 4K hybrid power wall. In the process of developing that system, the Rave RT-1250 became the company’s flagship workstation that you can now fit into your organization.
You can learn more about the Rave RT-1250 for yourself from the link over there. The landing page is newer than the RT-1250; data is being added as this goes out into the bit stream. Hit the Tech Specs for the lowdown on drives, USB ports and all that other neat stuff. My contact slipped me a copy of a new, almost-ready brochure that you should look out for. It has a lot of data on the NVIDIA technology used by the RT-1250 and why it’s important to you. But, in short, the Rave RT-1250 is something that you really should know about. It handles things like 8K graphics that are coming to you. Good stuff.
Thanks, Pal. – Lockwood
Anthony J. Lockwood
Editor at Large, Desktop Engineering