Who begs to differ..?http://www.techworld.com/opsys/news/index.cfm?NewsID=3477

That article is over a year and a half old, but does serve to illustrate why the semiconductor and high performance software application developers have been moving to multi core cpu's the last few years. billg

>That article is over a year and a half oldYou're right. I didn't notice. It was posted on a Dutch site as part of today's tech news..

But it's still pretty relevant to a lot of the discussions about FSX performance and whether or not future hardware without multi core processors will resolve the problem.billg

As said, an old article. But given that, it makes it even more incredulous that ACES made the development decisions they did.Ironically, ML is still just about applicable to GPU's, although it has taken an architecture change to provide the performance overhead potential.Allcott

It's an interesting problem for sure, and something I follow as an electronics engineer - while the article is old, the current limit for silicon based Integrated Circuit transistors is about 40nm (most production processors are about 65nm now I believe) - which means the prediction could well be true for Silicon devices.Still, back in the 90's they thought they couldn't go small enough to get past the 100MHZ limit on processors with silicon and they beat that barrier with a new fabrication process, so maybe they'll do it again, but we are getting close to the atomic limit for silicon transistors, no question about that - and it's smaller that makes faster (quicker slew rate).While there are alternative technologies being looked at (like chemical switches, that could be as small as 2nm), the fact is we are stuck with Silicon for the forseeable future, and thus multi-cores is the way we are headed.I love FSX, and it is annoying that ACES didn't make use of multi-core systems more, but my guess is reliable hardware wasn't available until too late into the development cycle, since FSX was 3 years in the making. But it's tough when our machines are all going multi-core, and we are stuck with a simulator that really is as fast as we can get a single core to go, and maybe for another 3 years until FSXI.Hopefully they will bring out some updates at some point that do help 'share the load' on multi-cores.

Reading about this, one gets the impression that the software is not even a variable. Like, how are going to run this Hummer at 45 miles a gallon? Of course, money is in the hard industry. That does not mean that performance and elegance is. Unless we have reached the ultimate in intelligence and elegance software-wise. Which is questionable at best.

True, however it's only a matter of time before we start seeing a surge in processing speeds that aren't related to multiple cores.Some genius out there will figure out how, even if it means going towards an entirely different technology. Hopefully this wall will expedite the move to different technologies (different processes other than semiconductor).James

>Who begs to differ..?>>http://www.techworld.com/opsys/news/index.cfm?NewsID=3477I don't think it's dead. It just stopped at Jack n the Boxfor a burger and fries. Think about it...The hardware and cpu bunch has to increase the performance fairly noticablyin order to keep hawking their products. They may be nearing thelimits of the present technology, but you can bet some junkfood eating engineer sitting in a smelly office littered withjumbo jack wrappers will find some way. You can pretty much count on it. MK