Re: Power supply efficiency on older model Sun equipment

Paul,

It really depends on what kind of fridge you have and what temperature 
you have it set on as well.  To assume "all" refrigerators consume the 
exact same amount of power is also a generalization and also quite 
faulty as well.  Your also assuming all E450's consume the same amount 
of power which is even more faulty than the refrigerator argument since 
depending on the number of power supplies, disk drives and other 
peripherals can make a pretty big difference in the power consumption.  
You also can't really measure the power consumption of any device since 
you can't measure the amount of heat it generates and how much more 
energy it consumes for your home A/C system to offset that heat.  As 
"scientific" as your tests are they are still generalizations and even 
tests done by an Engineering lab would be "relative" to the exact model 
refrigerator and exact configuration of computer system.   

If cost is really your issue why not eliminate the hardware completely 
and pay monthly for a Virtual Machine?  You eliminate hardware upgrade 
costs, power costs and administrative costs completely.  You also gain 
more bandwidth than your home cable modem and more reliability from the 
redundant power at the Colo that the VM lives at.  Oh not as fun and 
interesting you say?  My point exactly!

The real measure of efficiency is cost vs benefit.  For some people like 
myself having real Sun systems in a non production home lab environment 
allows us to be able to test theory and configuration on non production 
systems.  The hardware costs and energy costs are an investment in our 
careers.  A little bit of power to run these systems can pay of hugely 
in the right situation. 

Robert Chase

Paul Theodoropoulos wrote:
> Robert Chase wrote:
>> Hmmm,
>>
>> Interesting thread.  Not sure if I see the point though.  I have 
>> several Sun systems at my house including an Enterprise 450.  The 
>> problem that I have is not power it's cooling.  With all of my 
>> systems on I see an increase of perhaps $40 a month on my power bill 
>> however, my A/C system runs around the clock to keep the house cool 
>> if I keep all of them on.  On the larger more power hungry systems I 
>> only power them up when I need them such as the A3500FC fibre array 
>> and the Enterprise 450.  The smaller systems stay on 24/7.  I keep a 
>> Netra T1 two IBM x305's an IBM Netfinity 7600 a Foundry Edgeiron 
>> Switch and a UPS going 24/7.
>> http://www.systemv.org/site.html
>>
>> Even a big system like an E450 takes a lot less power than lets say 
>> an A/C unit or a Fridge or a washer and dryer.  If your looking for a 
>> power savings don't look at lightbulbs and computers look at your 
>> large appliances.  For a small Sun Workstation even an older model 
>> the power and heat consumption are probably about the same as a 
>> regular PC.  The difference between a Sun server and a PC server is 
>> probably too small to measure on most power bills.  What is 
>> misleading and difficult to measure is heat output which in turn ends 
>> up costing you money when your A/C unit keeps going 24/7.
> as the old saying goes, 'all generalizations are faulty...including 
> *this* one'.
>
> using the ever helpful Kill-A-Watt device, i've done plenty of 
> measurements around my house of how much power various things use. my 
> refrigerator is not much of a power consumer. over a 24 hour period, 
> it uses roughly 1.2 kilowatts. or in other words, about 60 watts 
> continuously. my home PC, a wildly overclocked core 2 duo with high 
> end graphics card, burns about 140 watts continuously, peaking at 
> about double that when gaming. the netra T1 out in the garage (also 
> overclocked, but just a little, and just for the fun of it) burns 85 
> watts continuously, and the D130 disk array attached to it burns the 
> same.
>
> the max power consumption of an E450 is rated at 1664 watts, though 
> obviously that with everything and the kitchen sink dropped into the 
> chassis. still, thats not trivial - it's more per hour than my fridge 
> uses in an entire day.
>
> and, a washer dryer, while they may consume quite a bit while they run 
> - they generally don't run 24x7, at least not in my house. i even 
> measured that - between two loads of wash and a dry (natural gas dry 
> though), they consumed 1kWh for what gets done only every several days.
>
> the lesson here i would say is don't guess - measure. the Kill-A-Watt 
> device is about $20 on newegg. well worth the price. when i discovered 
> that my bigscreen TV was burning 30 watts continuously while it was 
> 'off', i put it on a kill switch. there's lots of 'leakers' all around 
> a typical house. controlling those can save a lot on your energy bill. 
> my electric bill dropped by about 25% after i found and killed off all 
> the 'leakers'.
>