Hi Zane, Forgive me if I'm not using the right method to reply as I'm new to the list. I didn't see Al Kossow's post in the archive. I think that including a peripheral might make sense if it were a big hard drive but not if it were a printer. Anything that is involved in core computations for the computer (as opposed to input or output) would be fair game. I know that these computers do vary widely but I'm willing to go with what machines are available for metering. The big uncertainty is in getting measured power use of any kind. The exact boards that are installed are not as critical. Any ideas for folks who might have metered their PDPs? And the key thing to keep in mind is that the trends in performance swamp any errors in the measurement of power use, so I just need to get in the ballpark on power for the data to be useful. Defining system boundaries is important, so the peripherals question is one that we'd need to discuss. Thanks, Jon On Thu, 27 Aug 2009, Al Kossow wrote: I was wondering this myself. This is an interesting topic, but it might be difficult to define a "standard" configuration for these machines. Even the contents of the CPU boxes can vary widely. I suspect that many on this list are like me and tend to limit what what boards are in our CPU's to just what we need. How many people really need the CIS (Commercial Instruction Set) portion of the CPU in their PDP-11/44's or 3 Serial boards for terminals? I also limit the drives to what I need typically (I've even been known to unplug CD drives on PC's). Zane Hello all, I'm doing an analysis of trends in the energy efficiency of computation over time (for details, see the first report listed at http://www.intel.com/pressroom/kits/ecotech) . I'm using standardized performance measure from previous work by Bill Nordhaus at Yale, and then attaching measured power use to the machines for which Bill estimated performance. I got measured power data for the DEC PDP-1, PDP-4, and PDP-8 from Gordon Bell (now at Microsoft, who worked on these machines back in the day) but am seeking measured power use for some of the other machines listed below. I understand that some of these machines may not be unitary and so power measurements are more complicated. I'm interested in the sum of power consumption for the CPU, memory, and hard drives for typical configuration of the machines listed below. Ideally, this power measurement would be taken when the machine is running full bore, but I'll take what I can get. Does anyone have any measured power use handy? Thanks, Jon Koomey Power use Watts 1960 Digital PDP-1 2160 From Gordon Bell 1962 DEC PDP-4 1125 From Gordon Bell 1965 DEC PDP-8 780 From Gordon Bell 1968 DEC PDP-10 (KA10) 1971 DEC PDP-11/20 1972 DEC PDP-11/05 (11/10) 1972 DEC PDP-11/45 1972 DEC PDP-10 (KI10) 1973 DEC PDP-11/40 1975 DEC PDP-11/03 (LSI-11) 1975 DEC PDP-11/04 1975 DEC PDP-11/70 1975 DEC PDP-10 (KL10/DEC2060) 1976 DEC PDP-11/34 1976 DEC PDP-11/55 1977 DEC PDP-11/60 1978 DEC PDP-11/34C 1978 Digital VAX 780 1979 DEC VAX 11/780 1983 DEC VAX 11/730 1983 Digital VAX 750 1985 DEC VAX 11/750 1985 DEC VAX 11/780 1986 DEC VAX 8550 1987 DEC VAX 8350 1990 DEC VAX 6000-210 1990 DEC VAX 6000-460 1993 DEC 7000-610 1994 DEC 10000-610 1997 DEC 8400/5/350 x6 __________________________________ Jonathan G. Koomey, Ph.D. Project Scientist at Lawrence Berkeley National Laboratory Consulting Professor, Department of Civil and Environmental Engineering, Stanford University __________________________________ Jonathan G. Koomey, Ph.D. Project Scientist at Lawrence Berkeley National Laboratory Consulting Professor, Department of Civil and Environmental Engineering, Stanford University P.O. Box 20313 Oakland, CA 94620 (Please Email me to get a street address before mailing anything that has time urgency to it) 510-654-9634 Home and office 510-547-7860 Msg/fax 510-708-1970 Cell jgkoomey at stanford.edu numbers at numbersintoknowledge.com http://www.numbersintoknowledge.com http://www.analyticspress.com/ http://www.koomey.com