AMD Launches Low-Voltage Processors

mgoulding writes "CoolTechZone reports that AMD has released its low-power Athlon processors, which are designed to target the ultra-lightweight notebook market. The low-voltage chips will use smaller batteries and produce less heat. Acer plans to ship systems using the processors by the end of May." Acer plans to use them in their Ferrari line of thin laptops.

Comments

transmeta

[mandalayx]

The competitor that comes to my mind is Sharp and their Transmeta Efficeon processor. That notebook is quoted as being 2 pounds and 1GHZ/512MB/20GB/10.4" for $1499.

I do want a laptop for class and just having around, so I was strongly considering getting the Sharp, but constructive suggestions welcomed..

Re:transmeta

[Glock27]

The competitor that comes to my mind is Sharp and their Transmeta Efficeon processor. That notebook is quoted as being 2 pounds and 1GHZ/512MB/20GB/10.4" for $1499.

I don't think there's any comparison on performance...the 2800+ AMD part should completely smoke the Transmeta. Battery life...well I'm not sure but how long are you really planning on running unplugged at a time? I'd guess the AMD parts are intended to run for ~5 hours with moderate use.

Anyone have any benchmarks on the Efficeon?

Re:transmeta

[Anonymous Coward]

Transmeta Review

Re:transmeta

[Rodrin]

Well in terms of battery life the Efficeon blows the AMD away, however because of cache and optimization the AMD will blow a Transmeta processor away any day. The same goes for the fancy VIA processors that are "mobile". There is a difference between low power and mobile processors.

Re:transmeta

[Bryan Ischo]

This advice is baseless.

I have a generic 5+ year old Kapok notebook that is still going strong. It turned out to be a much, much, MUCH better value than any IBM, Toshiba, or Dell.

I don't think that laptop manufacturers differ so much that you should evaluate them based on brand. Evaluate them based on model and features, not brand.

Oh, except for Sony. STAY AWAY FROM SONY!

Re:transmeta

[evilviper]

I don't think that laptop manufacturers differ so much that you should evaluate them based on brand. Evaluate them based on model and features, not brand.

Not true at all... That is probably the single most important issue with notebooks.

Notebooks are not desktops. If a part is crap, you can't grab another and replace it... You have to go through the company. You better be buying from a company you REALLY trust, because notebooks are 100% lock-in.

Incidentally, I'm in the process of suing Sotec/Averatec...

Watt?

[Anonymous Coward]

I'm amped to get one of these!

the final frontier

[the arbiter]

I've been using nothing but AMD since 1998 and am an extremely satisfied customer.

If AMD has truly learned how to make a cool, low-voltage/low-wattage processor, well...all I can say is you might want to sell your Intel stock.

Go AMD!!!

Re:the final frontier

[tibike77]

Intel (as Microsoft and other "dinosaurs") will never get "really low in stockprice", because they have too much accumulated wealth to get out of most situations. Personally, I have used a mix of all kinds of MoBos and CPUs from everybody (yeah, even a Cyrix) and I can't say I have been extremely pleased nor displeased by any of them. The only news here is (therefore) that a new "cooler" (pun intended) CPU is here for the laptop market, hence prices will go down... shoppers rejoice :)

Re:the final frontier

[Rodrin]

Most geeks I talk to, unless they are crazy or otherwise weird, agree that AMD is the processor to go with. I can handle a little more heat when it comes to a $100 price difference most of the time. And now that technically the intel 64bit processors are a copy of AMD's Athlon64 perhaps AMD will lead and intel will walk behind. You never know.

Re:the final frontier

[geekoid]

well, IF AMD becomes the market leader, rest assure most geeks will switch to Intel.
These days, being a geek doesn't mean liking the best designed products, but you also have to like the underdog as well.

Re:the final frontier

[njcoder]

While lower voltage, cooler processors are great for laptops, it's the server market that could benefit most from these features. Paying $100 less for a processor is only one thing to consider when you have racks and racks of servers that you have to power and cool.

Re:the final frontier

[13Echo]

Not particularly. Geeks will normally just by the fastest CPU for the right price. For many years, that has been AMD. Intel is getting to be more competitive, but Athlons still perform better (in most respects) than a Pentium 4 of the same price.

There's a reason why AMD has slowly been gaining on Intel for market share, and that reason is why they accounted for 52% of desktop CPUs that shipped in a recent week. The "Intel Inside" campaign is wearing off, especially when Intel is trying to compete with AMD by releasing their own 64 bit CPUs that are based off of AMD's pioneering X86-64 chips. Who's doing the reverse engineering now?

Re:the final frontier

[Spoing]

1. well, IF AMD becomes the market leader, rest assure most geeks will switch to Intel.

   These days, being a geek doesn't mean liking the best designed products, but you also have to like the underdog as well.

Do you have a specific list to share (where popular is unfairly ragged on while obscure has an undeserved geek/wannabe-geek following)? Maybe I misunderstand.

For the record: I have bought an even split of AMD and Intel over the years. I don't see that changing, though it entirely depends on what goes around the processor. It's tough to pick a good sweet spot in laptops/notebooks mostly because of the extras (wanted and unwanted).

There are many gems out there that aren't popular so personally I get excited when I find them. Usually, they follow standards properly (in hardware and/or software). I also like booring and cheap generic basics too; mice, keyboards, hard drives, 100 or 100/1000 bit ethernet, USB 2 (not the dumbed down one), ... .

Re:the final frontier

[222]

Oddly enough the underdog also seems to have a better product at a lower price, and a much larger drive for innovation. Just my 2c ;)

Why is it called low-voltage?

[ForestGrump]

ok, it runs on a lower voltage.
But isn't the reality that they have a lower wattage?
Wattage being what really matters when it comes to power consumption and heat displacemnet.

-Grump

Re:Why is it called low-voltage?

[Rosco P. Coltrane]

Let's analyze and translate the above post shall we:

ok, it runs on a lower voltage.

That much is said in the article.

But isn't the reality that they have a lower wattage?

"wattage", also called power, only has to do with voltage for a fixed resistance. I suppose AMD lowered the voltage to lower the power consumption too, though, so what you said is obvious.

Wattage being what really matters when it comes to power consumption and heat displacemnet.

Yes, power is indeed what matters when it comes to power consumption.

In short: what a stupid post that was...

Re:Why is it called low-voltage?

[Humorously_Inept]

Voltage is a squared term in the general power function for a CPU. Lowering the voltage will have a significant effect on power consumption. As you might imagine, the chip's operating frequency is another term in the function.

Re:Why is it called low-voltage?

[Darth McBride]

IANAEE, but Ohm's law would suggest that a reduction in voltage would pull less current assuming the resistance is constant(V=IR). Power is equal to current times voltage (W=IV). Lower voltage leads to lower current which leads to lower power.

Re:Why is it called low-voltage?

[Anonymous Coward]

It's called "low voltage" because the term was coined by an engineer.

And it beat out the marketing department's entry "Less shocky stuff".

Re:Why is it called low-voltage?

[gerardrj]

Everyone, please take a step away from the keyboard and visit this site:

http://www.nmsea.org/Curriculum/Primer/How_is_el ec trical_energy_measured.htm

It's a very nice, plain english, tutorial about the relationship between Volts, Amps, Watts and Joules (the electric pantheon so to speak).

Re:Why is it called low-voltage?

[Kjella]

"wattage", also called power, only has to do with voltage for a fixed resistance.

No, wattage has to do with voltage times amperage. Resistance doesn't really eneter into it.

You forget Ohm's law: I=U/R, and so P = IU = U^2/R. So assuming you have a fixed R, wattage only depends on voltage.

In fact, it's basic electronics that, when lowering voltage, you have to increase amperage to maintain the same performance.

Unless you have the ability to change R, that is impossible. In itself less voltage = less power and less performance.

Lower power consumption means you can have higher transistor density though, which is what gives higher performance. I don't know what electronics class you went to, but I'd get my money back...

Kjella

tech info

[Coneasfast]

here is some tech info in case it gets /.ed

The other models - 2800+, 3000+, and 3200+ are rated to operate at 65W while the 2700+ is rated to operate at 35W. The Pentium M processor models - 1.5Ghz, 1.6Ghz, and 1.7Ghz are set to operate at 24.5W. The AMD 2700+ budget Mobile processor still consumes more power than the Centrino platform, which uses Intel's Pentium M chips. Another big difference between the rest of Athlon64 Mobile processors from AMD and the 2700+ budget level Mobile processor is the amount of L2 cache. The 2700+ processor has 512KB cache while the rest of the models house 1MB cache. The 2700+ is clocked at 1.6Ghz; the same clock speed as the 2800+, which has 1MB L2 cache.

Re:tech info

[Naffer]

The big problem that I see is that the CPU isn't responsible for all of the laptop's power consumption. The hard drive, GPU, chipset, RAM, and screen all use a ton of power too.
That said, the one thing that Intel really did right was that Pentium-M processor. I desperatly tried to convince my friend to get one, but instead he bought an Alienware with a 3.06Ghz desktop chip. (No speed throttling ability whatsoever) The result? 55 minutes of battery life on a single batter, just shy of two hours with the extra battery. Thats if you're just surfing the net.

Re:tech info

[number]

What you want are the Processor Electrical Specifications for any and all CPUs you can think of.

If you're serious about quiet (or preferably, silent) computing, the most valuable site I know of is Silent PC Review.

Re:tech info

[addaon]

The greatest power users in most laptops are, in decreasing order, cpu, screen, 802.11, gpu, hard drive, and other crap. Under normal usage, of course.

Re:tech info

[Brian Stretch]

What you have to remember is that the Athlon 64 series has very good power management, far better than Intel SpeedStep. When you're not doing anything stressful (say, reading Slashdot), the CPU slows and the core voltage drops down. The desktop Athlon 64's drop to 800MHz, my notebook says it's at 318MHz right now. If you need processing power that'll stomp a Pentium-M, it'll do that too. Oh, and you can run 64-bit Linux. And have the NX bit for neutralizing buffer overrun attacks in Windows XP SP2.

The heat output of my HP zv5000z Athlon 64 3200+ notebook with the CPU maxed out is fairly impressive but I've yet to be able to get it to overheat. (The Compaq r3000z is equivalent.) You can get 1920x1200 res screens for them now too. Pity HP used Linux-hostile Broadcom wireless cards and rigged their BIOS to explicitly reject non-HP wireless miniPCI cards (dumb!), but what can y'do (IBM and Dell do too, eMachines doesn't).

Anyhow, if I could get a mid-to-lightweight notebook with a low-voltage Athlon 64 2700+, top-of-the-line screen, Atheros WiFi (very Linux friendly), and GeForce 5700 Mobile, I'd probably upgrade again. The zv5000z's GeForce 440 is fine for getting work done and strategy games but woefully inadequate for UT2004. (Are you listening HP?)

Re:tech info

[dago]

Don't forget, AMD is rating the thermal power at maximum power drain, while Intel is rating them at typical usage, not maximum. So, it would be AMD_max = 35W vs Intel_avg = 25 W, and performance levels unknown.

And a nice thing about mobile amd cpu is that they are compatible with desktop boards, where intel are not in practice (400$ boards don't count).

Big Mistake, AMD...

[Deraj DeZine]

They should have gone the other way, making fast, but hot processors, and then marketed their notebook/travel iron combos.

No, I don't iron my clothes either, but it's the thought that counts, right?

Re:Big Mistake, AMD...

[Anonymous Coward]

Thanks for proving my theory that any attempt at a joke that gets posted early will be modded up.

Less heat?

[rms_nz]

If laptops start producing less heat, then what are we going to use for leg warmers on those cold winterery days?

Re:Less heat?

[dicepackage]

be careful there are 65,100 hits on Google about people burning their penis http://www.google.com/search?hl=en&lr=&ie=UTF-8&oe =UTF-8&q=%27laptop+penis%22&btnG=Search

I bet these OC nicely

[Anonymous Coward]

Generally, the less heat there is, the more room you have to try OC'ing. The current line of Mobile AMD processors has done very well.

Re:I bet these OC nicely

[INeededALogin]

Overclocking is just a bit over-rated.

What does over-clocking gain you? CPU Errors(instability), extra-heat, voided warranty, and possibly a dead CPU. Yes, you can come up with ingenious ways to keep that chip cool, but is the result really worth it especially when the extra speed you gain will be available in a non-overclocked model next month?

Re:I bet these OC nicely

[Carnildo]

Only if you do it wrong. If you know how to overclock, it gains you:

1) Serious savings. A low-speed CPU in a group can usually be overclocked to match the speed of the top CPU in that group without any special measures
2) Extra speed. You can almost always clock a CPU 5%-10% beyond the top speed for a processor group. If you're lucky, or using extreme cooling, you can get 20% or more.

And sometimes the extra speed isn't available for a while. When I purchased my current computer, the XP2100+ (266FSB) was the fastest CPU around. I got an XP2000+ (266FSB) and overclocked it to be an XP2200+ (333FSB), a chip that didn't show up until nine months later.

Re:I bet these OC nicely

[Humorously_Inept]

Actually, because of the way CPUs are binned to meet market demand, overclocking a CPU isn't necessarily doing anything to it that's bad. All CPUs are tested after fabrication and sold according to their capabilities. In the event that demand for a lower end processor is high, a processor that tests well can be marked as and sold as a lower-end model. It's still the same processor it was when it was tested, though. In this case, you might drop a CPU into your computer and experience exceptional overclocking potential. I have an XP1600+ that overclocks to 2200+ at default voltage and works ideally under a torturous FAH load. In fact, you could view the chip as being underclocked at the factory instead of overclocked by the user! :) Cooling is an important factor, but its nowhere near as important as good luck (or picking a good stepping).

They had to go and ruin it didn't they

[Wasteofspace]

Now what am I supposed to cook my eggs on????

Low Voltage Battery + Smaller battery = ?

[SCSi]

So this means that laptops will still have the same low battery life as they do now. Why dont they make low voltage processors with larger batteries so you can get more than a few hours of runtime.

Re:Low Voltage Battery + Smaller battery = ?

[Have Blue]

Strongly seconded... It's not like laptops themselves can get much smaller and still have usable screens and keyboards, so they may as well fill the remaining space with battery. I'd be much happier with a laptop that ran an hour longer than one that was a half-pound lighter.

Re:Low Voltage Battery + Smaller battery = ?

[Brandybuck]

Amen!

1998 == fat 3 hour battery
2001 == medium 3 hour battery
2004 == thin 3 hour battery

Have we become so sedentary that the weight of last year's laptop is going to cause heart attacks and strokes?

Great news for people like myself...

[proxima]

My next notebook will probably contain a low-power processor. There'e the Servelinux Enote for $800 that uses a Via processor like my mini-itx motherboard, but I suspect that AMD will be able to come up with something that's a little faster (it doesn't need to be blazing, but a 800 Mhz Via runs like a 600 Mhz P3 it seems).

I'd like to have either a 2.5-3lb subnotebook with a nice 12" screen (and preferably below $1k, like the Servelinux), or a ~4lb notebook that gets a much longer battery life than anything else on the market (besides maybe a Mac), but also is below $1k. No CDROM or large screen needed in my case, cause I'm not looking for a desktop replacement.

For now, though, the Servelinux enote is too obscure for me to look at it seriously, and I'll stick with my used 7020 (?) Toshiba Portege (at a little over 4 lbs I think, with a nearly useless battery).

I've personally seen and played with the enote, anyone have comments on other laptops in the same category (maybe from Transmeta instead?). Cheap, light, and fast, pick three; I like cheap and light.

Re:Great news for people like myself...

[proxima]

Cheap, light, and fast, pick three; I like cheap and light.

Whoops, I meant pick two of course.

Re:Great news for people like myself...

[RzUpAnmsCwrds]

"Cheap, light, and fast, pick three; I like cheap and light."

OK, then.

Go on eBay. You can get a Compaq Armada M300 (PIII 600, up to 384M of memory) for around $350. It's got a 11.5" XGA screen, it's 3.1lbs (including battery), and very thin and light (magnesium alloy case). It runs Windows XP and Linux fine, has a Cardbus slot for a wifi card, built in 56K and 10/100 ethernet, ATI Rage graphics with 4MB (horrible 3D, but fine for 2D), and a nice keyboard. Not to mention the fact that the power adaptor is tiny, and you can get a nice dock which only increases the thickness to around 1.2" (without changing the size otherwise) and adds a floppy + another drive (DVD-ROM+CD-RW, DVD burner, etc.).

hmm

[TechnologyX]

Ferrari line

But I'm on a Ford Tempo budget.. guess it's back to leg burns for me

What, for CMOS?

[The Mainframe]

"The low-voltage chips will use smaller batteries and produce less heat."

So these processors have built in batteries, eh? ;)

A non-overheating Italian car

[b0s0z0ku]

Acer plans to use them in their Ferrari line of thin laptops.
Is this the same laptop that was reviewed here a while ago? The one that makes revving sounds when booted up? Anyway, there's something wrong with putting a cool-running processor in a product named after an Italian car. Italian cars are supposed to overheat - when I had my FIAT/Pininfarina Spyder, part of the fun was sitting in traffic on a hot, humid summer day with the heater fully on to avoid cooking the engine.
b0s0z0ku

Re:A non-overheating Italian car

[dirkdidit]

I currently have a Fiat Spider and it's a zippy little Italian car. Part of the fun of driving it is knowing that you may not make it back in one piece. It just shows that you can have something that comes in a small package but is mean and loves to get hot, kinda like a midget.

Ferarri Line

[The_Rippa]

Great, tens of people will be enjoying the benefits of the new processor!

Why...

[rekoil]

can't anyone see the appeal of using chips like these in a ultra-quiet desktop model? A/V hobbyists would go nuts over them, providing that the CPU horsepower is sufficient...

Re:Why...

[murgee]

You can buy the Athlon XP Mobile chip now, and it works in a regular Athlon board. I'd imagine the same would be tru for the Athlon64 mobiles.. 'less you're talking about OEM machines.

Low Volt AMDs?

[An-Unnecessarily-Lon]

Man I want my 1.21 jigawatt proccesor.

hanging around

[poptones]

No way would I pay $1000 or more for a laptop computer. No matter what you do it's still going to have one of those cramped keyboards and limited expansion without carrying aroound a bag of USB goodies. And if you drop it then what? Yeah you can get an expensive new system with one of those "oops I dropped it" warranties, but that same money would buy a LOT of spare parts for an older, well built machine.

For $1500 I can have, like, 6 or 7 T600 stinkpads. Or maybe just one or two and a sack of batteries to keep it running 12 hours or more. Gets the job done, and if I drop it I don't care so much. "No one will ever need more than 500MHz and 512MB of memory."

I've had a lot of laptops and they all sucked compared to my (relatively) light, compact little thinkpad. They work great with linux and they feel so soft and squishy - if batman carried a laptop, he'd carry a classic thinkpad.

Re:hanging around

[Carnildo]

Cramped keyboards? The 15" PowerBook I'm typing this on has a keyboard with full-sized keys. They've got less travel than a normal keyboard, but that's the only difference.

Side by Side?

[randomErr]

Someone mentioned that Via's processors run slower then thier full powered conterparts. I can't wait until someone does a side-by=side of AMD/Via/Intel of the lower power chips.

Re:Side by Side?

[mandalayx]

*waves hand*

these are the benchmarks that you are looking for.

Source Bias?

[Anonymous Coward]

Looks like CoolTechZone is another FanBoi site (this one for AMD, or at least anti-Intel):

"AMD is currently leading the desktop processor market without a doubt with it's 64-bit processors."

Without a doubt? That sounds a little bit like "Ferrari is currently leading the automobile market without a doubt with it's Enzo model."

Yay, even hotter slow laptops!

[Anonymous Coward]

Soon we'll have laptops with 1THz 128-bit CPUs that can have up to 16 simultaneous hyperthreads running, all blocked on dog-slow laptop hard disk I/O 99.9999% of the time. We won't notice because our penises will explode in a cloud of steam as soon as we turn them on.

Awesome!

Re:Transmeta, also Side by Side?

[gabbarbhai]

Should they ever decide to actually ship their Efficeon chips in a laptop, it will also make an interesting comparison..

Low-voltage + PowerNow == PentiumM Competition

[david_reese]

Previously, only the VIA and Transmeta chips were in the same range of power consumption (actually much less) as the Pentium-M (which, tho I'm an AMD fan, is a pretty sweet processor). But those processors were pretty lightweight in the power department.

Now the Athlon64 mobiles, which already run at a cool 800Mhz when not taxed, combined with a voltage decrease, should create something that is at the same time powerful, yet battery-preserving.

Kicker: it's AMD64, so if you have 64bit OS and apps, it will really dominate.

Re:Low-voltage + PowerNow == PentiumM Competition

[raodin]

Actually they had Athlon XP-M low voltage models too, in the same range of power consumption as the new (Athlon64) ones. Fujitsu used them in their thin and light line, as well as Averatec. Probably a couple other thin and lights with them, as well. The summary for the article is a bit misleading, it makes it sound like these are the first LV cpus AMD has released. The actual news is that they're releasing LV Athlon64s.

Re:Low-voltage + PowerNow == PentiumM Competition

[evilviper]

Previously, only the VIA and Transmeta chips were in the same range of power consumption (actually much less)

Why do people keep doing this? Sure VIA processors are lower power than current Intel/AMD processors, but that's because VIA processors have terrible performance. If you compare a 1GHz VIA processor to a 500MHz Pentium, you'll see that both the performance and power usage are very similar.

I'm quite psyched about AMD's new processors though. AMD's XP procs have been beating Intel at maximum power consumption even with better performance, and AMD have finally started kicking their chipset manufacturers into line. I think the effect AMD64 had on AMD's whole company has been far more significant than the processor itself.

We Want Low Power CPU on the *DESKTOP* too

[Wolfier]

I'm sure a lot of us are looking towards the day where we can eliminate all crazy spinning fans from our computers.

Re:We Want Low Power CPU on the *DESKTOP* too

[Brian Stretch]

Or do the next best thing:

1) Buy an Athlon 64 PC
2) Enable PowerNow! power management
3) Buy a power supply with a variable-speed fan (I recommend this one) and enable CPU fan speed control on the motherboard (Q-Fan in ASUS's BIOS, IIRC).

When you're just reading Slashdot, the CPU runs at 800MHz and power consumption drops waaay down. When you're playing UT2004, the CPU runs flat out and the fans speed up. It works extremely well.

Re:We Want Low Power CPU on the *DESKTOP* too

[evilviper]

I'm sure a lot of us are looking towards the day where we can eliminate all crazy spinning fans from our computers.

Here we go again...
It always has been, and still is, entirely possible to have a system without any fans.

Just about any processor can be run with nothing but a large heatsink provide that you underclock it significantly. Drop a 2GHz processor down to 800MHz and you probably won't need a fan.

Oh, so you want a your system without a fan to be fast, eh? Well there's always water-cooling. But of course, you don't want to spend that much... Well, you can't have it all. If you choose silence, you have to spend a lot of money, or get poor performance.

Personally, I think the best way to just to replace the crappy fans and heatsinks with a tempurature controlled themaltake... Then when your processor is cool the fans will be slow and silent, but when your processor are running hot, the fans can cool everything down. I happen to like that tradeoff the best.

Choosing between AMD and Intel is difficult. Intel processor have a higher maximum tempurature, but they run cooler under typical load. AMD has recently been trying to fix this. This just happens to be a subject I detailed in my latest Journal entry.

Article and summary are wrong

[The Wicked Priest]

The announcement is of the mobile, low-voltage Athlon 64, not just the mobile, low-voltage Athlon -- which has been in the very machine I'm typing on for nearly a year. This is referenced in one place in the article, but the chips are misleadingly referred to simply as "Athlon" in the title, and several more times.

Junk

[poptones]

I've had two Vaios and they both died due to inadequate heat dissipation. Not enough cooling on the CPU = heat destroyed capacitors = flaky, dying laptops that cost more than the laptop is worth for repair.

Not to mention they came with the slowest hard drives of any preassembled computer I ever bought. I didn't pay for either of'em, and I'm glad. No way would I spend my pennies on ANYTHING labelled Sony.

Re:Laptops...

[jrockway]

Stastically, AMD isn't doing anything new really. However, in terms of mindshare, they're doing quite well. Remember the X86-64 spec? Intel seems to have copied AMD. This shows that AMD is ahead in something.

Frankly, whoever gives me a 5GHz 64-bit processor for $100 first wins in my mind :-D

Re:Not ohm's law

[rco3]

No. Ohm's law is in play. Charging a capacitor takes in energy, yes, but that energy is stored as charge, not radiated as heat. The heat comes from I^2*R loss. The I part comes when the capacitor must be charged.

So, it's really pretty simple. Lower voltages mean that the capacitors don't take in as much charge, and therefore don't require as much current to charge up. Less current == less I^2*R loss == less power consumed.

You can get similar improvements by reducing the size of the capacitances, which can be done by reducing the size of the FET gates... which means a smaller feature size.

You are correct that the transistors look mostly like capacitors, but you are incorrect in stating that power is stored in the junction and that creates heat. Power is dissipated in the interconnects, sources and drains, and in the vias between layers. This is also one big reason why we went to copper - lower resistance interconnects == less power lost to resistance.

Re:Laptops...

[Ahkorishaan]

Why does the mhz myth still exist? Intel will get there first I assure you. Who will be the better performer at the time? That is another question entirely.

Electrical theory tutorial...

[evilviper]

Lots of incorrect knowledge has been passing around /., so here's a nice quick tutorial about basic electricity.

Everything electric needs a certain number of watts to operate. Your Computer is about the same as a 100watt lightbulb. Wattage is voltage times current (amps), which means, less votage requires more current to do the same ammount of work.

Since voltage is not consumed, but current is, it only makes sense to use higher voltages, in order to preserve current. This can be seen very clearly on the specs of your computer's power supply. Your computer may need 100watts to operate, but you can choose between two different voltages. If you are running at 110v, you are drawing nearly 1 amp. However, if you throw the switch and plug in to 220v, you can operate the same equipment at under 1/2 amp.

I know this is a bit difficult to understand, so let's go with the shower analogy... Voltage is like water pressure, and current is like the volume of water. When you screw-on a "water conserving" shower nossle, what it does is increase the pressure. That makes the water spray just as far, while using less water to do it.

So, if you get the idea, you'll understand that decreasing a device's voltage does NOT mean you'll have better battery life. In fact, if all else remains equal, it guarantees worse battery life.

Now, I'd bet that AMD is decreasing the current used, while also decreasing the voltage, but that's just an asumption. The story only says it's decreasing voltage, which doesn't improve your battery life at all.

Re:Electrical theory tutorial...

[dpletche]

Unfortunately (for you), your grade school science lecture on electricity is not applicable to the power consumption of a CPU. There is no meaningful comparison to be made between a passive device like a light bulb or a water pump, and a nonlinear device like a CPU.

The dominant term in the CPU power consumption function is proportional to the square of the supply voltage, relating to the power consumed when charging a capacitor (or transistor). Using higher voltage can enable a CPU to achieve a greater clock rate, at the expense of diminished reliability and operating lifespan. However, since the CPU is not a system designed for the primary purpose of transforming electrical energy into heat, mechanical or radiant energy, it is not necessary to compensate for lower voltage with increased amperage in order to keep power level constant.

You are using a circular argument to support your flawed assumption that a CPU must maintain a certain level of power consumption regardless of applied voltage. Moreover, you are incorrect in stating that reducing voltage (assuming constant amperage) does not increase battery life. The battery is used to drive an efficient DC to DC converter in the power supply, so reduced supply voltage for the CPU translates into reduced battery current and power consumption.

Finally, your example about the water-saving showerhead is oversimplified, to put it nicely.

I wonder if I just fell for a troll...

Re:Laptops...

[servognome]

Why does the mhz myth still exist?
Because both AMD & Intel spent millions during the 200MHz-1.2Ghz years telling people clockspeed was everything. Its hard to deprogram people.