Eight cores on two processor packages or four dies and all of that in a single system. Needless to say that all components are standard shelf parts that can be purchased at any major computer part retailer. So what in the world has happened to good old computing as in emails and typewriter substitute?
The answer to that question is actually relatively simple, we are living more and more in the digital age, where content creation is taken for granted, whether it is good or bad content, doesn’t even matter, it is essentially the fact that it is available and, because of the ease of its creation, it becomes disposable. As somewhat of an oldtimer in the computing scene, it is hard to appreciate the fact that there is no more collection of files, everything is available somewhere somehow and even if the quality is a tad degraded, who cares, it is like reading a paperback instead of a hard-cover. Well, that gets a bit off topic but really what it comes down to is how fast can we create disposable digital media and does it matter whether we have to do it over again? With the system at hand, the answer is a clear no, just power it up, pay the electric bill and let her rip. In so far, the “Octopussy” system we reviewed here is possibly the penultimate cruncher for DVDs, MP3s and whatnot.
On the other side of the fence are the designers, using the same tools to create a digital reality in the first place in the form of CGIs. Time is money in that business and that alone justifies the use of extended numbers of cores, as long as the software supports them. Then there are the “other” designers, those who are running simulations of hardware designs, whether it is silicon or plastic or metal works or the simulation of thermals in complex designs of hybrid materials. Even though we did not run any of these simulations as benchmarks, this field is huge since machines are getting ever more complex and efficient at any task and it is a matter of time until the last human “idiot savant” who has a “feeling” for one or the other technology will have to submit to the growing power of electronic data crunching. The “Octopussy” is just another milestone on that path.
All of this said and done, during the time we were working on this article, an important decision was made at Intel to no longer really support the FBDIMM architecture except for use in ultra high-end servers. Power issues were the primarily cited factor in this decision. For my own part, I probably won’t miss them but if history has taught me something, it’ll be another 5 years and then we will see the next iteration of the same concept. This is not meant to say that the FBDIMM is fundamentally bad, it is just not good enough from a power and cost perspective to compete in a market where pennies on any item are escalated to multimillion $ cost over head or savings in the grand picture.
To finally come to the end of this article, it’s been fun working with this “Octopussy” and see what she can do. And that’s still what it all comes down to.
Source:Intel's V8 Media Creation Demo System -- Octopussy anyone?