I am very interested in processor technology, and how far can it go before the "chip" technology is maxed out. I would like to know what anyone else's thoughts on future electronic/digital technology might be. Where is the real big break through? Like the vacuum tube back in the day, or transistor in 1948 and the first integrated circuit in 1971. Those technologies paved the way for the most fantastic inventions man has ever seen, but all we see now is rehashed junk. Sure you might have a Quad Core running at 3.4Ghz but its still the same technology that was in use 40 years ago, just made to run faster. I want to know what is next! I guess there really isn't anything else to create for electronic equipment but silicon chips, or maybe our brains are just to pathetic to see the next path. I sit here and I keep thinking and thinking but I can't see what could be done to make the processor look like a punch card or gear.

Perhaps this would be the break through, just an idea but what would you do if you had an entire computer on on small chip, not just a processor but the RAM, Network Card, Video Card, Storage (Hard Drive), every thing that is currently added as separate devices. It would make the cpu case be as small as a MP3 player with just some plug in holes on the case for the attachments, but by then shouldn't need a keyboard if anyone can figure out a good speech recognition program that is. Perhaps a mouse and a screen. Like a laptop but better in other words, but its still a chip though, so even that is not really something different. Not the big break through. You can add more and more cores to a processor but its still a silicon chip. Basically to sum it up, I wonder will there be a device created that replaces the processor, like the transistor replaced the vacuum tube.

It's like RISC v CISC all over again.

/swoon

Why this over Cell?

Balancing supply and demand for all of the different permutations of CPU configurations could become pretty challenging. I wonder if AMD is going to be up to that...

AMD has to deliver something along the lines of Fusion soon. 2009 is too late.

Well too much technical stuff to read through but all i can say is it better beat Intel by a lot. Then Intel can come back and bet AMD by a lot. and so forth and so on...you get my drift.

I want more processing power and at lower cost damn it! Oh and less heat.

I like it. The new GPUs on these cores won't be able to have their own memory, I'll guess, but this will mean smaller sized motherboards for most office/laptop computers that are not geared towards gaming. I'd get one as a really small server. Of course, only after getting a powerful one :D

Wow, nice write-up Cyril, very nice. The company's plans are more interesting that I thought they'd be.

So let's see here: if all this pans out, I might be able to do one substantial upgrade before computing as we know it changes? Sheesh.

As soon as the notebook is plugged back into a power source, the discrete GPU will be switched on again, apparently without the need to reboot.

I bet that's a Vista-only feature (at least the ability to switch GPUs without a reboot). The guts of video drivers is not my area of expertise, but I don't think the XPDM would support that without some fairly nasty hacks in the driver.

"... A New Era"
Indeed. We now have OS's with 3d windows. This changes everything

So AMD as a gaming cpu is pretty much dead until 3rd-4th quarter 2007.... wtf is wrong with them...

Heterogeneous architectures comprised of smaller, specialized, cores has the potential to offer larger gains than a homogeneous architecture comprised of larger, general-purpose, processors. Not all algorithms are easy to parallelize (or parallelizable at all) so the scalability of cores, cores, cores, especially in the consumer market, is questionable.

ATI and AMD have been thinking in the long-run since the merger.

General purpose processor and specialized processors are converging.

Intel is going to follow suit as well.

Sounds like a fairly sensible development path to follow.The whole "more and more" cores scenario seems more like an arms race than efficient design.

Am I correct in thinking that AMD's approach of specialized processors on a single die is more sensible than Intel's apparent cores, cores CORES philosophy? I hadn't considered it before, but simply implementing a hell of a lot of cores does smack of the GHz wars.

I don't mean to demean Intel by this; a processor with 'hundreds of cores' is sure to present some nasty engineering hurdles, but I have to wonder if lower numbers of specialized processors is more sensible than high numbers of generic processors. Of course, given that Nvidia just switched to the latter, maybe that is the trend of the future.

Please Digg, if you will.

http://digg.com/hardware/AMD_thinks_there_s_no_future_for_the_multi...

Indeed, Fusion is a logical step for notebooks. If Intel can sell Centrino packages with processor + IGP chipset (and WiFi), integrating more of the chipset in the CPU itself makes sense.

I'm also excited about hybrid graphics, which is similar to what Sony currently implements in some of its notebooks, albeit not as transparently (i.e. requires reboot).