Fuel Cells are _much_ more efficient then a power plant. Power plants are only ~40% efficient. Fuel cells are ~85% (i believe) and the other 15% can be used to heat your home. As long as you get a good price on the fuel you should be able to create electricity for less then it costs the power company.
On macnn.com they have a link to an interview with Microsoft. In that interview they state that Office for Mac is their third most profitable product at just over $200 million a year.
The Mac might be a smaller market segment but it's still large enough for people to make money.
Very good explanation. I was wondering if someone would correct the previous post... Just thought I'd add a bit.
>>> There are three sources of power consumption in a cmos circuit. (1) leakage currents (very small), (2) charging and discharging the gate capacitors, and (3) so called "class A current."
2) is called the "switching time". There are different times for switching ON and OFF. Voltage however does effect switching time. A higher voltage produces a higher current that will empty/fill those caps faster. This lowers the switching time. The power saved from a lower switching time is offset by the added power drawn from running at a higher voltage.
Lots of funny things now start to happen. You raise the voltage to raise the current. The added current creates heat that increases the resistance that lowers the current. Basically, voltage and switching time aren't linear. This is why it's most efficient to run at low voltages.
Everyone should note, the only time a cpu uses power is when it is in the process of changing states (every clock cycle). (Ok, the only _relevant_ power. Better?) If you double the frequency you double the number of switches hence doubling the power consumption. All this assumes the voltage remains constant.
>>> The second item on the list, charging/discharging current, is quite important.
Let me upgrade that to _most_ important. It's the switching time -- the only relevant time the cpu draws power.
>>> Electromigration pales in comparison to soldering irons, drills, and other tools of the trade.
Ha ha ha... Obviously spoken from experience. I'll second that notion!!
>>>as those gates expand, they start passing more current >>>MOS has a negative tempco (unlike BJTs). They conduct LESS current when they are hot. That's why you don't need to worry about thermal runaway when playing with power mosfets.
Very true. That's one of the reasons you overclock at a higher voltage -- to help push those electrons through the extra resistance. (Thus producing more heat, thus requiring more voltage, thus....;)
>> More current = more heat >> No. P = IV, you know that.
Look at the CPU as being a black box. You give it x volts it draws y amps and produces xy watts. Now if it draws more power (ie, more current at the same voltage) it will produce more heat.
>> Current is what causes heat, NOT voltage. >> No. See above.
Yes but the voltage is what determines the current. The current then determines the temperature that determines the resistance. The increased resistance then usually reduces the current.
Ohhh, lets just say this... Current causes heat. Voltage determines the current. You could say that voltage causes heat but that isn't exactly right.
Back to the main topic. If you just overclock the CPU you're very unlikely to cause any damage. Raising the voltage is what puts you in the danger zone. To avoid damage, run at the lowest possible voltage.
Actually sounds like a good idea. Are their any other comparable standards out there? If apple gets ATI and NVida (spelling?) on their side it might actually become a standard.
Of course you still have the problem of USB having to be on the Video card. How about a USB input plug on the vid-card? That might do it if motherboard manufacturers start including USB ports accessable from the motherboard. (Yes, I know AT motherboards are like this but they don't count.) Then just a small cable would be required.
Probably just a heatsink. Peltier coolers are too expensive and would draw a pile of power from the already small power supply.
Whay would be cool is to use a peltier on the new G4e when they come out this winter. With an on-board cache it should overclock much like the Celeron. Apple will probably ship these with only a heatsink so if a peltier were added they should overclock quite well.
Current G4s don't overclock that well due to the external cache. Oh well, I don't have one.
I think Apple should create an opensource project to address this issue. Well maybe not all opensource. I'm sure Apple has some decoding stuff that it fiercely protects. But if they could put together a project where almost everything is opensourced and Apple just puts is the required pieces that they don't want public... voila, quicktime on Linux. If Apple provides pre-compiled binaries of their stuff the Linux community should still be able to customize the player to their liking.
The way I see it (and I'm not a programmer... yet) the hardest part of a Linux version would be the X11 based code. The decoding routines shouldn't be any different then on the mac or windows. I say let the Linux community write the X11 stuff. After all, they know it better then Apple.
Ahh, all the issues that would arise from this plan. It'll never happen. Just a thought that came to me a few days ago. Wanted to share.;)
I beg to differ. More speed does equal more heat. Look at the Alpha. Designed for speed and can be used to heat your apartment in the winter!!
The powerPC was designed to use little power. Sure it's fast but it would be even faster if they didn't compromise on the design and let it use more power. Just look at IBM's POWER series of chips. They draw lots of power and are much more powerful.
Slashdot Mirror
Fuel Cells are _much_ more efficient then a power plant. Power plants are only ~40% efficient. Fuel cells are ~85% (i believe) and the other 15% can be used to heat your home. As long as you get a good price on the fuel you should be able to create electricity for less then it costs the power company.
G4s never got a speed bump. The notebooks will probably use a G4e because it's better on power then the G4.
On macnn.com they have a link to an interview with Microsoft. In that interview they state that Office for Mac is their third most profitable product at just over $200 million a year.
The Mac might be a smaller market segment but it's still large enough for people to make money.
Very good explanation. I was wondering if someone would correct the previous post... Just thought I'd add a bit.
;)
>>> There are three sources of power consumption in a cmos circuit. (1) leakage currents (very small), (2) charging and discharging the gate capacitors, and (3) so called "class A current."
2) is called the "switching time". There are different times for switching ON and OFF. Voltage however does effect switching time. A higher voltage produces a higher current that will empty/fill those caps faster. This lowers the switching time. The power saved from a lower switching time is offset by the added power drawn from running at a higher voltage.
Lots of funny things now start to happen. You raise the voltage to raise the current. The added current creates heat that increases the resistance that lowers the current. Basically, voltage and switching time aren't linear. This is why it's most efficient to run at low voltages.
Everyone should note, the only time a cpu uses power is when it is in the process of changing states (every clock cycle). (Ok, the only _relevant_ power. Better?) If you double the frequency you double the number of switches hence doubling the power consumption. All this assumes the voltage remains constant.
>>> The second item on the list, charging/discharging current, is quite important.
Let me upgrade that to _most_ important. It's the switching time -- the only relevant time the cpu draws power.
>>> Electromigration pales in comparison to soldering irons, drills, and other tools of the trade.
Ha ha ha... Obviously spoken from experience. I'll second that notion!!
>>>as those gates expand, they start passing more current
>>>MOS has a negative tempco (unlike BJTs). They conduct LESS current when they are hot. That's why you don't need to worry about thermal runaway when playing with power mosfets.
Very true. That's one of the reasons you overclock at a higher voltage -- to help push those electrons through the extra resistance. (Thus producing more heat, thus requiring more voltage, thus....
>> More current = more heat
>> No. P = IV, you know that.
Look at the CPU as being a black box. You give it x volts it draws y amps and produces xy watts. Now if it draws more power (ie, more current at the same voltage) it will produce more heat.
>> Current is what causes heat, NOT voltage.
>> No. See above.
Yes but the voltage is what determines the current. The current then determines the temperature that determines the resistance. The increased resistance then usually reduces the current.
Ohhh, lets just say this... Current causes heat. Voltage determines the current. You could say that voltage causes heat but that isn't exactly right.
Back to the main topic. If you just overclock the CPU you're very unlikely to cause any damage. Raising the voltage is what puts you in the danger zone. To avoid damage, run at the lowest possible voltage.
Radeon should fit. Might need a bigger heatsink to get rid of the fan but other then that it'll be the same size as the Rage128.
Actually sounds like a good idea. Are their any other comparable standards out there? If apple gets ATI and NVida (spelling?) on their side it might actually become a standard.
Of course you still have the problem of USB having to be on the Video card. How about a USB input plug on the vid-card? That might do it if motherboard manufacturers start including USB ports accessable from the motherboard. (Yes, I know AT motherboards are like this but they don't count.) Then just a small cable would be required.
Probably just a heatsink. Peltier coolers are too expensive and would draw a pile of power from the already small power supply.
Whay would be cool is to use a peltier on the new G4e when they come out this winter. With an on-board cache it should overclock much like the Celeron. Apple will probably ship these with only a heatsink so if a peltier were added they should overclock quite well.
Current G4s don't overclock that well due to the external cache. Oh well, I don't have one.
I think Apple should create an opensource project to address this issue. Well maybe not all opensource. I'm sure Apple has some decoding stuff that it fiercely protects. But if they could put together a project where almost everything is opensourced and Apple just puts is the required pieces that they don't want public... voila, quicktime on Linux. If Apple provides pre-compiled binaries of their stuff the Linux community should still be able to customize the player to their liking.
;)
The way I see it (and I'm not a programmer... yet) the hardest part of a Linux version would be the X11 based code. The decoding routines shouldn't be any different then on the mac or windows. I say let the Linux community write the X11 stuff. After all, they know it better then Apple.
Ahh, all the issues that would arise from this plan. It'll never happen. Just a thought that came to me a few days ago. Wanted to share.
I beg to differ. More speed does equal more heat. Look at the Alpha. Designed for speed and can be used to heat your apartment in the winter!!
The powerPC was designed to use little power. Sure it's fast but it would be even faster if they didn't compromise on the design and let it use more power. Just look at IBM's POWER series of chips. They draw lots of power and are much more powerful.
Hey, don't be to hard on the guy. The images on Apple's site were probably made with photoshop. They look to good to be snapshots.