2020

[programmer76]

**computers will run at 5.2 GHz speeds in the year 2020 becuase right now computers run at 3.4 GHz **

[Ancient Dragon]

Here's an interesting article about how computer speed has changed since the 1980s. That's not to say that processor speed will continue to increase at that rate, most likely motherboards will just support more cores, for example the processor in my computer has 8 cores, and I know of motherboards that can handle 2 processors which would give it 16 cores. That's an awful lot just for me to play a few games :)

[Warrens80]

True true

[pixelsoul]

**computers will run at 5.2 GHz speeds in the year 2020 becuase right now computers run at 3.4 GHz **

I only partially agree with this, because it will be sooner than the year 2020 that this will be happening. They already have and AMD processor that runs at 5ghz, and Intel is working on one now that they say will be faster than 5ghz.

Here is an interesting related article
http://www.zdnet.com/amd-beats-intel-by-unveiling-worlds-first-5ghz-processor-7000016677/

http://www.extremetech.com/computing/170023-amd-vs-intel-the-ultimate-gaming-showdown-5ghz-fx-9590-vs-i7-4960x

[Warrens80]

Could be. That is if we don't have a complete shut down

[Deep Modi]

yeah,
it may more than it so...
i think so

[Warrens80]

Yes it will be so

[ddanbe]

Who will tell if we will use computers in 2020, the way we do now?
Perhaps we all talk via our smartphone, to quatumcomputers living in a cloud.
Who had a smartphone or a phablet, 7 years ago?

[Warrens80]

I believe in the sometime near future there will be a huge electronically shut down in the us.

[Ancient Dragon]

shut down in the us.

Probably world-wide event. Knock out the sattalites and all communications and computers as we know it today will stop.

[mike_2000_17]

Knock out the satellites and all communications

Easier said than done, there are far so many satellites that a large number of satellites could be "knocked out" without really compromising the whole system. Global telecommunication systems are massively redundant by design, there aren't really any single points of failure in it. And also, how exactly can you knock out a satellite? Satellites on geosync orbits (where most tele-com sats are, except of a few russian sats on high elliptical orbits) are very far away (about 36,000 km altitude), are not affected by problems like space debris, are too far to be destroyed remotely, have safety low-gain antennas in case attitude control is lost, and are designed to run for decades before needing to be replaced. They are not about to disappear, I can tell you that much. Even GPS satellites are quite safe, even though they are much closer to the Earth. The more dangerous zones for satellites are in low-Earth orbits (LEO), where observation satellites are (weather sats, spy sats, etc.).

The Internet is not about to go down either, after all, it was designed for redundancy and failure safety in the first place (i.e., military networks relayed through planes, helicopters, vehicles, and bases, which are all subject to suddenly disappearing). The only thing that really threatens the internet is people with economic motive to compromise it's inherent openness and neutrality, which are the key aspects of its fail-safety and redundancies.

computers will run at 5.2 GHz speeds in the year 2020

As pointed out, I think they already do surpass the 5GHz mark. Several years back, a friend of mine had an overclocked, liquid-cooled 4.5GHz computer.

In any case, we are starting to reach certain physical limits that prevent much further improvements on metrics like clock speed and transistor density.

For instance, I think state-of-the-art CPUs these days have 32nm to 22nm transistor sizes, and the theoretical limit (with current silicon-based technology) is not too far away from that (5-10nm), which is due to the size of silicon crystals, probabilities of malfunction if too few carriers are present, and quantum effects that occur at such small scales. So, notwithstanding a big technology shift, we have kind of reached the limits of "miniaturization".

As for clock speeds, they are limited by the switching power of transistors and heat evacuation capacity. Heat evacuation has some real hard limits that have been reached already, and can only be extended with liquid coolant (instead of air (fans)) or full-blown refrigeration cycles (i.e., heat pumps), neither of which are particularly convenient in your home computer or laptop.

And transistor switching power is also difficult to affect by much, because it's inherent to the transistor's physical principles (i.e., the time it takes for electrons to travel, which is about 1/3 the speed of light). Btw, transistor power consumption is due to the fact that a momentary short-circuit is created every time a transistor switches, and thus, the faster the switch the lower the power, but the speed is limited by the speed that the carriers can move into the gate (carriers: electrons or "holes" (absence of electrons, i.e., positive charge)).

This is why a lot of the focus lately has been on making smarter CPUs. Like multiple levels of cache, smarter pre-fetching / branch-prediction / out-of-order execution, multiple cores, parallel modules, etc... These are all designed to make the CPUs faster (or allow programs to do more) without actually increasing the clock speed or the overall number of transistors / modules on the chip itself.

And in that vein, I do think that computer technology still has a lot to give, and I don't think we've reached the end of Moore's Law just yet. We will ride this wave of "smarter CPUs" for a little while longer, and then, I think there will be a paradigm shift pretty soon. It could take the form of memristor technology, quantum computing, or neurosynaptic chips, who knows?

[Warrens80]

Idk man.

[Ancient Dragon]

In any case, we are starting to reach certain physical limits that prevent much further improvements

I seem to recall those same comments about computers in the 1980s.

[pixelsoul]

I seem to recall those same comments about computers in the 1980s.

And before that even, the Abacus!

[Warrens80]

Think about it. What if all the power house countries went to war with each other and we all sent EMPS at each other

[Ancient Dragon]

By EMPS do you mean this?

[pixelsoul]

I need to create an account to read that article?? No thank you!

[Ancient Dragon]

I don't have an account at that site and I don't have any problems reading it.

[mike_2000_17]

What if all the power house countries went to war with each other and we all sent EMPS at each other

You watch too many movies. You are not alone though, even the US Air Force had to make a video "Hollywood vs EMPs" to dispell the hollywood myths about the power and impact of EMPs. For one, the power required to send a real wide-spread EMP is only achievable with nuclear weapons. And if you were to generate enough EMPs to seriously cripple the entire power grid or tele-comm systems, it's not the outages you should be worried about but the nuclear fallouts. On a more practical level, EMPs can be used to induce glitches on targeted systems, possibly temporarily shutting them down or damaging some electronics components. But the effect is very limited (either very narrowly targeted, or wide-spread with very little impact except a few minor glitches). It is nothing like in the movies, that's called fiction.

[Ancient Dragon]

According to the article whose link I posted

A single nuclear weapon detonated at an altitude of 400 kilometers over the United States would project an EMP field over the entire country, as well as parts of Canada and Mexico.

But you are right -- we'd have many more problems to worry about than EMP if that ever happened.

[mike_2000_17]

A single nuclear weapon detonated at an altitude of 400 kilometers over the United States would project an EMP field over the entire country, as well as parts of Canada and Mexico.

That's exactly the kinds of things people should not say, because it lacks any kind of measure of impact. This statement is trivial, i.e., you could potentially create a large enough nuclear weapon such that if it detonates in the sky above Texas, you would be able to measure / detect an EMP in Vancouver, Canada. But the important questions here are "how big would that nuclear weapon need to be?" and "how strong will the impact of the EMP be when it reaches Vancouver?". The reality is, to have an EMP impact in Vancouver that would be strong enough to shutdown just the most vulnerable electronic equipments, you would probably need a nuclear weapon so powerful that it would likely turn Texas and Mexico and every state in that radius into a massive crater.

To have any real impact, the EMP needs to carry a lot of energy (eletro-magnetic energy), and in a blast radius, energy diffuses in an cubic rate, meaning that you very quickly end up with very little energy left in the pulse. And the larger the radius of effect, the larger the initial pulse needs to be, by a cubic power.

A more likely scenario to try and cripple the entire USA through an "EMP attack" would have to involve detonating thousands of nuclear weapons in a sort of massive grid pattern that would cover all the US. That's very unlikely, and the effects would be much more than just a temporary outage, more like turning the entire place into a massive nuclear wasteland.

[Ancient Dragon]

I don't know -- that article was based on some congressional report

Based on eight years of research and analysis, 50 years of data from nuclear tests and EMP simulators, and never-before-attempted EMP tests, the commission found that any nuclear weapon, even a low-yield one, could potentially pose a catastrophic EMP threat to the United States, mainly because of the great fragility of the electric grid.

So, are you saying that's just a lot of bullshit??

[mike_2000_17]

What I'm opposed to is the lack of quantification of anything. When it says "could potentially pose a catastrophic EMP threat", there's no measure in that statement: "potentially" could mean anywhere from 99% chance to one chance in one trillion; and, "catastrophic" could mean anything from wiping out all of the human race to causing significant damage to the infrastructure (require significant repairs, and possibly an extended down-time). This kind of alarmist language with no real measure or proportions irritates me to the highest degree.

From the 2010 report I glanced over (which was rather poor quality, from my judgement as an engineer who has studied those kinds of effects), even they seem to conclude that there are serious concerns about preserving the equipment and avoiding wide-spread power failure due to a cascade of failed equipment. But it seems limited to bursting relays and burned PLCs, and they do a lot of wild speculations about the likelihood of equipment to be damaged. And there are many questionable methods in their study, like, for instance, using upset-limit voltages to decide if a pulse of a certain voltage will damage them, without considering the power density. Also, they seem to model all plugged-in electrical equipment as being a large antenna (the cable) plugged into the equipment, without modeling ground capacitance and dynamic response of the equipment and transmission equipment. Long story short, it does not convince me that there is very much of a threat except for a few mishaps that would be quite easy to recover from (maybe a temporary, local outage, at most). And again, the report does not "quantify" the threat because they admit themselves that there is far too much speculation involved to really quantify it, but they say that the yield of an "EMP attack" is likely very small and insignificant. If a nuclear attack is perpetrated, they won't blow the thing in the air (to get an EMP), but rather blow it on the ground (to destroy a city), because if you get your hands on a nuke and want to do damage, that's the only logical thing to do.

[ddanbe]

This kind of alarmist language with no real measure or proportions irritates me to the highest degree.

Me too, newspaper are sometimes filled with it. Guess they think it sells more newspapers . . .