No Smaller Silicon

Regarding Yabullshithoo, that sounds like FUD tactics.

#49: If someone could actually get it right, I think everyone would be much better off. Also, does anyone know of a free Win98 memory recovery / monitoring program like MemTurbo? Thanks.

P.S. I love how today\'s programs leave little turds in your registry all over the place.

#49 Lex. I really like your idea and I think it is in development at least in part. I saw a news program about six months ago that interviewed a few R & D ppl at Microsoft. Basically they were developing this stupid looking parrot that sat on your desktop and observed your actions. It would then try to anticipate what you were going to do next and often give verbal suggestions. If it hought you were going to load a certain file it would have it ready... things like that. At the time they were targeting windows 2000 for this but that has not happened to my knowledge. It was definately an example of thinking outside the \'box\' and quite innovative I thought. i think your idea, however, is much more broad in scope and has a lot of potential. Find a coder see and see what you guys come up with. :)

-Bits-

#51, wouldn\'t 1000 mhz = 1 ghz?

Thanks!

-Lex

Here\'s another angle:

In the infinitely customizable world of computing, why can\'t someone develop a \'bot\' program so us users can dial in our systems and tweak our cpu\'s functionality to whatever we want, within certain parameters. Not overclocking, but operating within design-specifications and maintaining it.

For example, we could use a bot program that activates and then hides in the OS background. This bot would monitor the cpu (or any compliant hardware for that matter) and actually learn about the functionality of the cpu and it\'s daily uses. For a student in medicine, the cpu might perform specific daily tasks, while a home business user may have completely different needs. In raw terms, it would be like taking a bot from Q3 or UT and assigning it the job of learning which files are opened the most, which resources or settings are selected when those apps are used, etc. If we can develop a bot in UT or Q3 that can hunt your ass down and frag you, can\'t we have a similar bot (or AI) perform more vital functions? This bot would be very different from just a defrag or scan disc app, because the bot would always be active and eventually learn when the user was active and when it would be okay to shunt primary cpu power to system upkeep. I\'m talking about having this bot go online for you and spend the time updating, retreiving email, transmitting email. It would be a bot that takes several days - even weeks to get to know just how it\'s user is actually using the computer, then going for maximum tweakage.

Imagine how simple (and relieving) it would be knowing your system was functioning at maximum effeciency, and knowing the all of the behind-the-scenes tweaking would be happening in real-time, without interfering with your daily life. I mean, there are times when I am not using my system to it\'s fullest potential. Like when my system resources drop to 71%. This bot would activate at a certain point and restore system resources to max, without me having to suffer with complete system bog-down, or reboot, or pull my f\'ing hair out.

My refrigerator handles a very valuable task, and it maintains it\'s own operating environment without my day-to-day intervention. My telephone performs it\'s tasks without my needing to become involved at all.

Shouldn\'t we be thinking outside of the *box*?

Thanks!

-Lex

It might not be a bad thing that intel is held back a bit because of a technological brick wall. If they keep making very similar chips for a long time they can get the manufacturing costs down which means cheaper chips for us. Plus it promotes the idea of smp as a general standard which I personnaly think is long overdue. It could also force intel to expand their product line a bit. One thing that would speed up a computer as much as any chip advances (well almost) is advances in harddrive design. In the past few years we have gotten some nice little increases (udma for one) but nothing mind boggling. Fast solutions (like ultra wide scsi and arrays and firewire) are still a bit pricy. How about a solid state drive that the average person can afford? That would be a major performance increase. Your swap file could theoretically be as fast as your ram (that is if they redo some mb architecture to take advantage of it). I know I am dreaming a bit here but think about the possibilities. All I am saying is that the cpu is not the only place to make performance adavances. Tha glaring performance dog in the average home computer in my oppinion is the harddrive. They make them bigger all the time (that sells) but not any faster.

What\'s this shit I keep hearing about nano-machines? Aren\'t there companies, or at least universities, working on them? There has to be a link between building nano-machines and linking hundreds if not thousands of them together to solve this cpu problem. Maybe nano-machines could build the cpus.

Anyone know anything valid about this stuff?

Thanks!

-Lex

Better architectures means we can get faster computers with even less clockspeeds than today\'s chips..

For example, Steve\'s Geo Metro and a BMW M-3 (heh) both use sheet metal.. but there\'s a slight speed difference?

Current chips already have many layers, but with true 3d processors, there are unresolved heat dissipation issues, just to name one problem.

Noononono, no we start using 3d chips, with more than one layer, because if the chips get too small, electron tunneling will take place, so instead of a flat chip, they make one like a block, etc.

This might not work for palm pilots, etc.

Dear God, Intel is actually going to figure out how to innovate for themselves again.

#32, Intel has stuff alot higher than 1100mhz, you just can\'t buy them. I think they\'ve got up to P7\'s over there. I\'m not sure if this is true, but I do know someone who works there, and that\'s what he claims.

cool, anyone has an old towercase for my notebook?

It is rare but it does happen.

This is a load.

Aside from alternative technologies such as RSFQ superconducting logic and single electron transistors, they still have some headroom to go in the current technology.

I met a guy the other day at UCBerkeley making 25nm feature size CMOS circuits (thats .025 micron, or 1/10 the size of current tech.) They are making small circuits on the order of 100 transistors, but they work. The problem is finding a fast way to make them this small, not that if you make them this small they wont work.

The only current wany to make chips this small is with direct write electron beam lithography. This is very slow as it has to scan over every detail, instead of doing it all at once like standard lithography.

cot

The answer to a dilemma like this is probably to take the road AMD has with their k7. Have more happen each cycle instead of increasing the Mhz, thereby slowing down the need to rise higher in Mhz speeds. This is why the k7 is faster than the intel chips, it can do more in parallel. Now I know this is limited because in the end the need is always there for raw speed but this is definitely something companies will consider, re-engineering their processors to do more per cycle, more in parallel.

-cre3519]
http://www.belizeinfo.com/k4
PS- please dont give me any more shit, I just want to have some intelligent conversation, and my opinions are just my opinions, I am not forcing you to comment on them.

Intel is trying to bolster fear in the cpu market by releasing official information that we have hit the technical \'wall\'. Then in 4 months or so, they bust out a breakthrough technology that no one else has. It would make them look like heroes, and re-gravitate all cpu purchases back to Intel again.

If Intel seriously looked at their architecture, they would see it\'s not the \'final solution\'. They are entering a time when they are going to be spending bucks on R&D, which must be compensated with consumer interest in their products, which should turn into sales growth.

Why else would Intel release a statement that ultimately says, \"Sorry. We can\'t go any further.\" That statement, without some sort of \'breakthrough advancement\' only leads to fear, and fear leads to dark side.

Thanks!

-Lex

Yeah, sCary, what the fuck.
It\'s taking so long people have started
to have intelligent conversation.
Hurry it up.

Scientists say that every few years, but then new technology shows up that allows them to keep going. I\'m sure it\'ll happen this time, too.

well i read in in some magazines that they have this new thing to make chips out of its closly related to diamonds but, is realy cheap. Anyway, it can sustain very high temps and still work so that means we dont even need to make the chips smaller.

Wasn\'t there an article on ZDNet recently which says that scientists may have found a way to create molecular transistors which are no wider than a few (maybe a hundred?) atoms in diameter, thus permitting CPUs to continue to be improved?

I can\'t find the link on ZDNet.... their retarded advanced search doesn\'t even perform the \"order by date\" they say it does.