gamerk316 :
Now that we are moving to 20GB bandwith caps, that will take care of itself in short order. I fully expect bandwith caps to shrink the internet by 20% or more, and with a lot less bandwith consuming sites.
20 GB per month is about 650 MB per day or 7.8 KB/sec 24/7. The Flash, Javascript, and ad-heavy "Web 2.0" stuff won't come close to that figure. Neither will e-mail or the amount of HTTP/FTP file downloading most people do. Streaming music won't either, unless you stream 128 kbps music for more than 12 hours a day. What will go over those limits in a hurry is watching a lot of online video, videoconferencing, and peer-to-peer applications. The cable ISPs *hate* online video and want to quash that, while the DSL ISPs *hate* any sort of online voice/video communication and want to quash that. Both hate P2P due to the fact that they'd have to upgrade their networks to really deal with it and they don't really like being bothered by the MAFIAA lawyers.
And for the record, embedded systems have had multithreading support for multiple CPU's for ages; the system I work with, all I have to do is one command in software for 85%+ scaling.
Basically any OS except for Windows and MacOS has supported SMP for quite some time.
We have reached a point though, where diminishing returns will cause us to hit a brick wall with what we can do grapically (which I personally hope will force companies to improve game physics in the meantime).[/quotemsg]
Reynod :
/goes back to listening to the matresses flooping ...
I don't even want to know
joefriday :
Bah...overkill! Sempron LE-1100 at 850MHz/0.8volt ftw. It runs my Media Center PC (although it will clock up if it CPU load is >80%)
Bah, that's all overkill. My laptop has a C2D T7250 running at 600 MHz (100x6) at idle, down from the stock 800 MHz (8x100.) All SpeedStep-aware P6-class mobile chips I've encountered have a lower multiplier of 6x and the GM45 chipset drops the FSB down to 100 MHz yet for some reason the default idle on this chip is 8x100 rather than 6x100. I simply adjusted the EIST settings to correct this. However, Intel has locked the idle Vcore (Vcc_SLFM) at 0.850 volts, so I can't undervolt it any