RE: Forcasts, why won't anyone believe them?
The assumption is wrong. A server motherboard and CPU draws the same power regardless of what it's doing. Lap-tops and the like are different and you actually pay extra for that design. Most circuts these days are nmos technology. Only in cmos does the power go up with the frequency. Peripheral usage, like disk drives, are also constant since the largest power draw goes to keeping them spinning. The seek mechnics are trivial. Floppy drives and cd-rom drives are different. But, most servers do not keep those spinning constantly. Ergo, for all intents and purposes, servers are a constant power draw. They can be rated. But, isn't this a topic for the DataCenter list?
-----Original Message----- From: Jeff Cours [mailto:jeff@ultradns.com] Sent: Tuesday, January 16, 2001 4:18 PM To: nanog@merit.edu Subject: Re: Forcasts, why won't anyone believe them?
Sean Donelan wrote:
One thing that would help.
Sun, Dell, Cisco, Compaq, Juniper, etc. Can you please start listing the true power draw of your equipment, not just the fuse rating. It would make forcasting a lot easier, if we knew ahead of time how much the equipment will really draw.
I'm not sure they can. Doesn't the actual power draw of a piece of equipment depend on what it's doing? For example, a rack full of Pentium III's that are acting as routers are mostly doing integer calculations, running bus transceivers, and driving communications links. That same rack full of Pentium III's acting as a render farm for your favorite Hollywood movie will be doing floating point intensive calculations, wide-spread memory access, spinning the disk drives, and, because of the extra heat, working any variable-speed cooling fans harder. I'd expect a measurably higher current draw in the second case.
It might be possible to come up with some sort of average power draw, but Electrical Engineers really hate to give out numbers like that because people base their designs on them instead of on the worst case power draw, and then when something fries the EE winds up getting the blame. That's why most engineering disciplines derate components and allow a safety margin, which I suspect is where the fuse rating comes from.
- Jeff
-- Jeff Cours Senior Engineer UltraDNS, Inc
Once upon a time, Roeland Meyer <rmeyer@mhsc.com> said:
The assumption is wrong. A server motherboard and CPU draws the same power regardless of what it's doing. Lap-tops and the like are different and you
Not necessarily. For example, when a server is idle, Linux "halts" the CPU between interrupts (instead of doing a "busy wait" loop). This significantly drops power usage and heat generated, which can also mean lessfan power. So, a busy Linux server generates more heat and uses more power than an idle Linux server. I think FreeBSD does the same; I don't know about other OSs. -- Chris Adams <cmadams@hiwaay.net> Systems and Network Administrator - HiWAAY Internet Services I don't speak for anybody but myself - that's enough trouble.
Yes, but is it *really* measurable? I mean, are we talking a difference of more than a watt (in a machine that already pulls 80 to 120 watts)? 1% difference in power consumption could be related to someone passing wind in the building on the same power grid.
So, a busy Linux server generates more heat and uses more power than an idle Linux server.
I think FreeBSD does the same; I don't know about other OSs. -- Chris Adams <cmadams@hiwaay.net> Systems and Network Administrator - HiWAAY Internet Services I don't speak for anybody but myself - that's enough trouble.
Mac OS [9|X] has very fine grain control over the CPU speed. This is utilized on laptops for enhanced battery life, and on desktops for lower power use and heat generation. For a real example, the digital thermometer sitting on top of my G4 cube shows about 84 F while it idles. When I launch dnetc to start cracking RC5 keys, the air temp above the grill (no fan on this machine) rises to 97 F within about 5 minutes. Since I'm plugging the Mac, I might as well also point out that the latest 773MHz G4s use 17 Watts max load. I'm often dissapointed that Apple doesn't do more to tout the energy efficiency of their hardware. -pmb At 11:26 PM -0500 1/16/01, Alex wrote:
Yes, but is it *really* measurable?
I mean, are we talking a difference of more than a watt (in a machine that already pulls 80 to 120 watts)?
1% difference in power consumption could be related to someone passing wind in the building on the same power grid.
So, a busy Linux server generates more heat and uses more power than an idle Linux server.
I think FreeBSD does the same; I don't know about other OSs. -- Chris Adams <cmadams@hiwaay.net> Systems and Network Administrator - HiWAAY Internet Services I don't speak for anybody but myself - that's enough trouble.
-- Ring around the Internet, | Peter Bierman <pmb@sfgoth.com> Packet with a bit not set | http://www.sfgoth.com/pmb/ SYN ACK SYN ACK, |"Nobody realizes that some people expend We all go down. -A. Stern | tremendous energy merely to be normal."-Al Camus
On Tue, 16 Jan 2001 23:54:40 PST, Peter Bierman <pmb@mac.com> said:
Since I'm plugging the Mac, I might as well also point out that the latest 773MHz G4s use 17 Watts max load. I'm often dissapointed that Apple doesn't do more to tout the energy efficiency of their hardware.
Hmm.. 17 watts. Of course, that's just the CPU, but I just noticed that where I'm working right now is primarily lit by 2 22-watt flourescent lights.. So.. How many of you kill the flourescents in your back-office cubicle farms during a California-style power crunch? ;) For that matter - send the cubicle critters home. You ever *looked* at how many watts the average human body outputs? ;) Valdis Kletnieks Operating Systems Analyst Virginia Tech
Valdis.Kletnieks@vt.edu wrote:
On Tue, 16 Jan 2001 23:54:40 PST, Peter Bierman <pmb@mac.com> said:
Since I'm plugging the Mac, I might as well also point out that the latest 773MHz G4s use 17 Watts max load. I'm often dissapointed that Apple doesn't do more to tout the energy efficiency of their hardware.
Hmm.. 17 watts. Of course, that's just the CPU, but I just noticed that
The CPU chips on the previous G4s use only about 4 or 5 watts (I suspect it hasn't changed much). It's a lot less than a P3 because it's RISC vs CISC chips.
where I'm working right now is primarily lit by 2 22-watt flourescent lights..
So.. How many of you kill the flourescents in your back-office cubicle farms during a California-style power crunch? ;)
For that matter - send the cubicle critters home. You ever *looked* at how many watts the average human body outputs? ;)
Valdis Kletnieks Operating Systems Analyst Virginia Tech
Most people I know don't power their bodies from a wall socket... Of course, people that spec A/C do indeed take into account the heat generated by the occupants. Regards Marshall Eubanks Multicast Technologies, Inc. 10301 Democracy Lane, Suite 201 Fairfax, Virginia 22030 Phone : 703-293-9624 Fax : 703-293-9609 e-mail : tme@on-the-i.com http://www.on-the-i.com
From: Roeland Meyer <rmeyer@mhsc.com>
The assumption is wrong. A server motherboard and CPU draws the same power regardless of what it's doing. Lap-tops and the like are different and you actually pay extra for that design. Most circuts these days are nmos technology. Only in cmos does the power go up with the frequency. Peripheral
Nope. Most are CMOS and the power goes up for both NMOS and CMOS with increasing clock frequency. That may not change your conclusion though.
usage, like disk drives, are also constant since the largest power draw goes to keeping them spinning. The seek mechnics are trivial. Floppy drives and cd-rom drives are different. But, most servers do not keep those spinning constantly. Ergo, for all intents and purposes, servers are a constant power draw. They can be rated.
Chris
:Nope. Most are CMOS and the power goes up for both NMOS and CMOS :with increasing clock frequency. Let's not split any more hairs. xMOS power consumtion is tiny and irrelevant to the original context of this thread. -brian
From: Roeland Meyer <rmeyer@mhsc.com> Subject: RE: Forcasts, why won't anyone believe them? Date: Tue, 16 Jan 2001 19:49:45 -0800
The assumption is wrong. A server motherboard and CPU draws the same power regardless of what it's doing. ...
I am just a software guy, but I believe that most _systems_ have a large current draw when they are powered up, (e.g., all those fans, some of which are pretty large). I believe that the original comment was observing that there is often quite a difference between the rated (peak) load offered by a device and the average load. Both are useful to know. (I assume that everybody, sometime in their career, has learned about the difference between peak and average load when they tried to power up all of their routers at once and popped the breakers.) -tjs
participants (9)
-
Alex -
Brian Wallingford -
Chris Adams -
Chris Beggy -
Marshall Eubanks -
Peter Bierman -
Roeland Meyer -
Timothy J. Salo -
Valdis.Kletnieks＠vt.edu