Slashdot Mirror
Why We Need a Second Moore's Law
Roland Piquepaille writes "In its April issue, Wired Magazine argues that we need a second Moore's law, this time about overall efficiencies of our computers and other electronic devices. The subtitle of the article summarizes it: "If we don't do something about increasing battery life, we're toast." Michael S. Malone, the author, says that the first Moore's law is endangered, not because the semiconductor industry cannot build new generation of chips, but because we will not be able to provide them with enough power. And he contends that the problem arises from the fact that we are using more and more wireless devices, which obviously are not connected to a plug. This overview contains selected excerpts of this eye-opening article."
What we need is a fourth axis of development - a systematic improvement of overall system efficiency, from the individual silicon gate, through motherboards and displays, all the way up to the Internet itself. How do we do it? Exhaustively.
Exactly. When processor speeds and memory was low the industry did their best to fit what they could in the limited space. Now that we have more room we are being lazy and only concentrating on making things "larger than life" instead of faster and smaller.
We should really start to concentrate on making the software run best under what we currently have. I know that Intel and Kingston wouldn't exactly be happy but our pockets and our grid would.
Moore's law is an observation, not something that the industry is forced to follow. You can't just say "we need more efficiency - let's define a new Moore's law".
Phil
I dont think theres need for new laws, even for Moores law, we just need more technological advancement and new innovations ...
This sounds a little dodgy to me. This statement seems to imply that a law is 'needed' to fix a current problem (i.e. batteries not keeping up with processor power). But why would some contrived 'law' do anything to solve this problem? After all, the original Moore's law was a prediction - no more, no less. No one has ever actually been guided by it.
I feel that putting the problem forth in this way is just clouding the issue.
I hear there's rumors on the Slashdots
What many people don't realize is that Moore didn't just say the one famous sentence. He wrote a short paper explaining his predictions and they were far more complicated than simply doubling power while maintaining cost. He qualified it by explaining it would only be possible if certain things happen. He was well aware of certain limits which we've now passed with unexpected technologies.
So while many here will complain his prediction was flawed because he didn't consider so many other things, remember he actually had a lot more in mind than just regularly doubling speed.
Developers: We can use your help.
I think the battery power does not have to be solved by only "internal" system efficiency, but also by "external" system efficiency.
What if the places to charge our devices become pervasive, and just like you get can find a gas station almost everywhere you seem to be running out of gas, you should be able to find a place to charge your batteries.
Of course this is easeir said than done. The "external" system is developed well for vehicles running on gas - but it is not well developed for vehicles on electric power. That is why electric cars lag so far behind ....
Anyway, the crux of my post is that the system efficiencies not have to improve internally in the "super"devices, but also externally to the devices. .
To see a world in a grain of sand, and then to step back and see the beach where the sand lies
n its April issue, Wired Magazine argues that we need a second Moore's law, this time about overall efficiencies of our computers and other electronic devices. The subtitle of the article summarizes it: "If we don't do something about increasing battery life, we're toast."
I can imagine the board room at Intel where the chairman is yelling, "The 3rd quarter numbers suggest we aren't going to make Moore's law this year! I want people to double their efforts -- cancel lunch until further notice!"
I can guarantee that if wired magazine invents a new moore's law, it is going to have zero effect on technology. Anyways, Moore's law is based on an observation, maybe we should look at the growth of power requirements and fit it to that.
I suggest we call the wired law: Moron's Law
For batteries to get better at a Moore's law rate, we need some different physical laws. But we can, as other posters mention, improve on efficiency of other parts. Cooler-running processors and low-power wireless - a la BlueTooth or 802.11[whatever-letter-means-low-power] will help.
To build a better battery has been a goal for longer than computers have been in around.
OK, when on the move, how many people who are doing word processing need more than the features of WordPerfect 5, the early versions of Excel for Windows and that kind of thing?
What we need is a really low electrical power CPU - optimized to take as little electricity as possible, but which is capable of running these kinds of applications acceptably quickly. It probably doesn't need to be more than 50MHz. Put this in a ultra-lightweight laptop style case, using solid state storage for disk (you can get USB memory sticks with 512MB which is more than sufficient for this class of computing) and have the battery go a day or two between charges.
My mobile phone is a case in point. Although it's not a word processor, I've got an organizer, email client, lightweight web browser, camera, SSH client, IRC client and pager all rolled into one, and it'll go ten days without a charge on standby, and can be used for 7 hours on one charge with a tiny battery. I can even make phone calls on it. Make essentially a notebook with mobile phone technology, and you've got an excellent portable internet terminal that you can write documents, make spreadsheets, compile small programs etc. on.
Oolite: Elite-like game. For Mac, Linux and Windows
I second that. It makes me sad when people say that it doesn't matter the software is inefficient, because computers will get faster, which will solve the problem. What this means is that, because the developers were sloppy, the users have to pay more (they need faster and typically more power-hungry machines).
Making more efficient software benefits users _now_, instead of in 5 years when computers have gotten faster and new power sources have been invented, and new software will have been invented that needs even faster computers. Having a lot of CPU power is no excuse for wasting it.
Please correct me if I got my facts wrong.
"I'm thinking more of a way to utilize kinetic energy and translate it into stored power."
The problem is not with the abiliuty to generate the power but a way to efficiently store it with something that is at a reasonable size. Batteries/Power Cells are not moving at a very fast pace compared to the rest of the industry. We can generate all the power in the world but we don't have a small cost effective way to store it yet.
Evolution or ID?
As a whole, we're generally a pretty environmentally-conscious bunch. That said, we geeks find ourselves dependent on more and more powerful, long-lasting batteries that do horiffic damage to the environment when not properly disposed of.
How then do we balance our concern for the environment against our ever-increasing portable power needs? For the time being, these seem to be conflicting goals.
Socialism: A feeling of discontent and resentment caused by a desire for the possessions or qualities of another.
... but what are batteries made of? What is out there of their components that can be improved or replaced so they will last longer?
... "green" batteries?
And also forgive my ignorance of biology (I was an average and disinterested student in high school),
but some plants use the sun, and store energy in fruits, right? Isn't there some model we can use to make, I don't know
I can't count how many times I heard a professor say "don't optimize", "memory is cheap".
When everyone is more worried about making thier code pretty instead of efficient, well we get what we've got. Feh.Free Mac Mini Yeah, it's
Hasn't it been pretty well proven that too much manual optimisation doesn't pay off? The time taken to optimise delays entry to the market, causing the optimised product to be obsoleted by newer (unoptimised) technology.
Isn't this pointing to a requirement for better automated design software, able to do optimisation in essentialy zero time. Any optimisation between manufacturers will require their design tools to automatically exchange data. I can't see too many manufacturers being prepared to swap such detailed design information (unless they are 'open source').
I disagree with the terms of this article. He's right if you're talking about putting the fastest possible technology in portable devices -- but we've NEVER been able to do that. If you want a "Moore's Law" for portable devices, it's that when you drop the clock frequency of a processor its power requirements drop still faster. The processors in current cell phones put the original IBM PC to shame, and in ten years they'll be approaching the speeds of current desktop machines, and even though those desktops will be light years ahead, the portables will be plenty nice.
Or we could simply go nuclear. The biggest problem with radioisotope generators is the problem of keeping them out of landfills, and eventually our water supply. If we solve that problem, we can power laptops and cellphones for half a decade or better.
Javascript + Nintendo DSi = DSiCade
Is the law still working for he amount of power used? I understand that the newest chips do use more power, but shouldn't that be the approach.
Good point, but there are other factors at work. What the article ignores is that the current law is still working fine. The amount of power used by an x-MHz chips has been steadily declining. I remember when a machine with a 16 MHz processor needed a 200W power supply -- now we have machines that are orders of magnitude faster and still use a 200W power supply.
What has changed is that people insist on having desktop power in a laptop form factor. Another change has been a move to architectures that do more per clock cycle -- old processors often took 4 or more cycles per instruction. New processors, using pipelined and superscalar architectures, dispatch and retire multiple instructions per clock cycle. Technologies like hyperthreading boosts utilization of the CPU's logical units and thus increase power. Modern CPUs draw more power because they do more per cycle.
Moore's Law is fine. That mobile devices lag desktop devices in performance (or suffer from poor battery life) has nothing to do with Moore. The computing power of a 1 watt processor continues its steayd rise.
Two wrongs don't make a right, but three lefts do.
The author is just completely and totally wrong... There's no better way to put it, I'm afraid.
Let's start off with an example... My good old 33MHz 486 notebook had a fairly large battery that would only last about 2 hours. Now, I've got a 1.2GHz notebook, with a far larger screen, smaller/lighter battery, and the battery life is much BETTER, not worse. What's more, there are notebooks much faster than mine, with 5+ hour battery life, and are still lighter.
The main reason is that ALL the components are getting more effecient. The hard drive is a significant drain on your batteries, but they are getting quite a bit more effecient every day. Things like the LCD backlight are becomming much more effecient (and brighter) at the same time. But that's just to start...
Power supplies are getting MUCH more effecient, and batteries are improving quite a bit as well (not quite doubling every 18 months, more like every 24 months).
Although the author seems to think otherwise, processors are becomming more effecient as well. My notebook only uses 30watts at MAX CPU/HDD utilization, and averages about 15 watts. Desktop processors are becomming more effecient quite quickly, just not as quickly as the speed is ramped up. While a 500MHz AMD processor used 42W, a similar 1,000MHz processor used 65W. That's right, effeciency IS improving quickly.
But that's only on the desktop front. If you look at notebooks, you will see that effeciency is even closer to matching performance improvements. It's just a matter that Intel/AMD are willing to spend the extra money on making notebook chips more effecient, while they aren't willing to spend much money on making desktop chips use less power. (which is why I'd personally like to have an ATX mobo that accepts a mobile Intel/AMD processors).
Slashdot gets worse every day... Pipedot: News for nerds, without the corporate slant
Don't bother arguing speed, saying that the powerbook is years behind in MHz, etc. The powerbook is just better optimized to use less power and run longer.
Don't worry, I stopped arguing with mac users back in 1997. I could mention that AMD and Intel also have low-power cpus, but I won't. I could also find a few laptop that beat the powerbook in battery life and processing power, but I won't.
That's like this friend I used to have. When we had arguments, it was never a matter of finding the best answer, it was a matter of him justifying *his* answer, and the argument would never stop before you agreed with him or just stopped talking to him.