Do I believe Gaming Cards and Workstation cards will eventually become one and the same-? Well, I have to leave that up to the experts in silicone modeling. What I said, is that I discovered the clear and present gap between the two; but after reading your article, considered the gap could possible narrow, in time. (Never say Never). For today, however, I have merely clarified the distinctions between the two. Thanks for helping me to see and better understand the issues with improved clarity.
A common mod is turning a gaming graphics card into a workstation one through firmware modding and driver forcing. Though, the end result is usually not as efficient. You believe gaming and workstation cards will become one and the same?
That would be great. Changing modes depending on the application!
Recently, while customizing a new system for my business, I was making ala' carte selections from HP's build-your-system website, and the only GPU Cards available were all gamer-based processors. Trying to identify workstation GPUs, none were even offered in the machines I was attempting to customize. Obviously, the gaming market has a higher market potential for HP than does the engineering design market.
I ended up going with a different manufacturer, and building a system with an NVIDIA Quadro K2000M, coupled with the Intel Core i7-3610QM Processor --- much like what you are describing in your test set ups in the article.
So, you taught me a couple new things; mainly, it looks like the gap may be closing between the two separate markets. But as you concluded, at a higher price. At the time of my system configuration ( Q3-2012) the NVIDEA k2000M was a $250 upgrade from the standard GPU offered in the computing package. Comparatively, how is the new Fireproof R5000 priced-?
The limitation of only one user being able to utilize the GPU at a given time is a significant drawback, in my opinion. In comparison, Nvidia's forthcoming GRID vGPU technology will allow businesses to deliver the performance needed for engineering, design and video applications to multiple employees at any given instance. This will be accomplished through a desktop virtualization infrastructure, with a lesser number of servers at the back-end than you would otherwise require without GRID vGPU. Hence, relatively speaking, it should also take care of the other drawback i.e. cost.
For industrial control applications, or even a simple assembly line, that machine can go almost 24/7 without a break. But what happens when the task is a little more complex? That’s where the “smart” machine would come in. The smart machine is one that has some simple (or complex in some cases) processing capability to be able to adapt to changing conditions. Such machines are suited for a host of applications, including automotive, aerospace, defense, medical, computers and electronics, telecommunications, consumer goods, and so on. This discussion will examine what’s possible with smart machines, and what tradeoffs need to be made to implement such a solution.