Timeline for Why do CPUs need so much current?
Current License: CC BY-SA 3.0
26 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Apr 16, 2018 at 3:35 | history | edited | alex.forencich | edited tags | |
| Sep 21, 2016 at 15:52 | comment | added | user56384 | @AndrewMorton How to build a shunt ammeter widget (USA): as I recall it is 3 feet of #16 wire folded up (not coiled) and bridging the white side of a duplex outlet with the ear broken off. 12 milliohms. 1000 W reads 0.1 AC Volts. | |
| Sep 21, 2016 at 3:50 | comment | added | Stack Exchange Broke The Law | "why can't a CPU operate at 1 V and 1 A?" - they can! A CPU can even operate at a few volts and a few milliamps! But not this particular one. This CPU is designed for speed, not low power consumption. | |
| Sep 21, 2016 at 0:38 | comment | added | Stack Exchange Broke The Law | @EricLippert That is not correct. The current through all parts of a series circuit is equal, but the CPU is not in series with the wall supply. (Also, you say "all parts of a system"; the power LED is part of the system, but are there 64 amps going through it? Or the CPU heatsink? Or the little screw that holds my graphics card in?) | |
| Sep 20, 2016 at 18:10 | comment | added | Eric Lippert | @AndrewMorton: It would certainly be wise to use an ammeter designed for that task, yes! Jeff explains it here: blog.codinghorror.com/why-estimate-when-you-can-measure | |
| Sep 20, 2016 at 17:44 | comment | added | Andrew Morton | @EricLippert Also, "Put an ammeter on your computer plug" without carefully describing how to use an adequately rated meter to do so is a dangerous suggestion: a naïve user might put the ammeter across the live and neutral terminals of the mains plug, leading to a not-so-fabulous adventure to a hospital accident and emergency department. | |
| Sep 20, 2016 at 14:33 | comment | added | Ben Voigt | @dim You are quite correct, but I'm not up to the task of describing a buck converter in the limited space allowed for comments using terms someone without an EE degree would understand. | |
| Sep 20, 2016 at 14:20 | comment | added | dim | @BenVoigt That's the description of a charge pump voltage divider. Typical SMPS supplies used for computers operate completely differently: AC-DC (in the supply block) use flyback transformers, and internal low voltage DC-DC conversions (on the motherboard) uses buck topologies (with an inductor). Your description is right, but it is not what is actually used here. | |
| Sep 20, 2016 at 14:03 | comment | added | Ben Voigt | @EricLippert: Consider this: a group of 10 capacitors is placed in series, and charged at 1A to 10V each by a >100V supply. Now, transistors between the capacitors switch, and they discharge in parallel, each giving 1A at 10V, for a total of 10A at 10V. The discharge time is then equal to the charge time. Have a second set that charges while the first discharges, and vice versa. DC-DC conversion has a lot of additional complexities to it, but this is the basic concept. | |
| Sep 20, 2016 at 14:02 | answer | added | Dmitry Grigoryev | timeline score: 17 | |
| Sep 20, 2016 at 13:49 | comment | added | alex.forencich | @EricLippert The motherboard in your computer contains a multiphase DC to DC converter that steps the supply voltage (12V in the case of a desktop, probably 12-19V in the case of a laptop) down to the core supply voltage. This is done with constant POWER, so the output current ends up being 10-20 times the input current. Not to mention the 12V supply in a desktop computer also comes from a switching power supply which also converts with constant power. The CPU in your computer probably has at least 100 power and ground pins to handle the current. | |
| Sep 20, 2016 at 12:45 | comment | added | MSalters | The conserved quantity is energy, and on average also power. If a CPU draws 64 Watt, then the power supply must draw at least 64 Watt from the socket. That's <1A even at 110V. | |
| Sep 20, 2016 at 12:10 | comment | added | Andrew Morton | @EricLippert "it would have to be pulling 64 amps out of the wall" - I have a suspicion that the CPU would not be operating on 110 V. | |
| Sep 20, 2016 at 11:48 | comment | added | Eric Lippert | The total current through all parts of a system is a constant; if a CPU was using 64 amps off the power supply then it would have to be pulling 64 amps out of the wall, but my computer is plugged into a circuit protected by a 15 amp breaker. And there is certainly no way my laptop batteries are providing 64 amps. There's something wrong with your math. Put an ammeter on your computer plug and you'll soon know its amperage. | |
| S Sep 20, 2016 at 9:47 | history | suggested | user | CC BY-SA 3.0 | Better title |
| Sep 20, 2016 at 9:37 | review | Suggested edits | |||
| S Sep 20, 2016 at 9:47 | |||||
| Sep 20, 2016 at 9:33 | history | tweeted | twitter.com/StackElectronix/status/778165136628613120 | ||
| Sep 20, 2016 at 7:10 | comment | added | Uwe | If the clock is slow and the CPU is a smaller one, it can operate at 1 W or even less. | |
| Sep 20, 2016 at 5:45 | history | edited | scanny | CC BY-SA 3.0 | Improve English usage, capitalization, units formatting |
| Sep 20, 2016 at 1:30 | vote | accept | Leonardo Bosquett | ||
| Sep 19, 2016 at 20:07 | comment | added | PlasmaHH | In certain areas of the CPU, leakage/static currents dominate, like in the 70nm era I have seen 70% of the power consumption of L1/L2 cache being cited as static. | |
| Sep 19, 2016 at 20:02 | comment | added | glen_geek | Do a FinFET transistor count on a modern CPU. Not every FET conducts current from Vdd to ground, but even so, 64 A gets distributed over *a very large number * of these switching FETs. | |
| Sep 19, 2016 at 20:02 | answer | added | alex.forencich | timeline score: 73 | |
| Sep 19, 2016 at 20:01 | comment | added | user98663 | Modern CPUs (none of which are 'simple') require multiple voltage rails all with their own power requirements. Your question makes many assumptions and has many erroneous statements. You must consider all power requirements and not just those for a single rail. | |
| Sep 19, 2016 at 19:55 | review | First posts | |||
| Sep 20, 2016 at 11:05 | |||||
| Sep 19, 2016 at 19:50 | history | asked | Leonardo Bosquett | CC BY-SA 3.0 |