> "but it is a nice arguement for upgrading our systems to faster ones "

> Yea, but then what do you do with the old 'puter? Before you know it you will
> have it setup and running BOINC also.....I know, been there, done that. Now
> I'm running 8 with two of them dual cpu boxes, so in all I'm running 10 cpu's.
> Of course in the winter they are the only heat I need in the comuter room so
> not all is totally lost.
>
Jim

LOL. Spread em out and call them central heating. Probably more efficient..

This all reminds me actually. I need to get a new fan. Sounds like the bearings are a little off and makes a racket whenever it changes speed.... Mind you the extra friction adds to the heat.


cRunchy

One way to cut down power usage if running seti (or most other BOINC enabled projects) 24/7 is to run it in a ramdrive, have all your harddrives spindown when they are not used and have enough RAM to avoid swapping. Don't run any other programs that may use the drives during the night etc - this will have the drives spinup a lot and shorten their expected lifetime.
Having the drives spindown will also decrease the amount of heat they produce.

Furthermore run those boxes that you don't have a monitor for without graphics card (most modern grfx cards use a lot of power) - instead use a networked thing like tightvnc.

Make sure that you have no extra fans that are just there to make noise and look cool. These actually consume a little power as well. If possible you should make the airflow of your case better, use passive cooling for the CPU and the (nonexistant in some systems) grfx card and only have the PSU pull the hot air out of the top of the case.

Buy a better PSU - preferably one of those new ones that transform the power much more effeciently, making less heat and thereby wasting less power.
As it was mentioned earlier in the thread you should buy one that has a _larger_ wattage than what your system consumes...

I guess that's it for now =)
Happy crunching!

> One way to cut down power usage if running seti (or most other BOINC enabled
> projects) 24/7 is to run it in a ramdrive, have all your harddrives spindown
> when they are not used and have enough RAM to avoid swapping.

A couple of Watts isn't really worth worrying about IMHO.

---

Grant
Darwin NT

> Also, i'm more concerned about the health of my PC, running at 100% usage
> 24/7. Anyone have any idea of the life expectancy of an Intel P4 HT 3.2G?

Of quality components & properly assembled & maintained it'll last longer than it will be usefull for.

---

Grant
Darwin NT

> $335,311.20 X 12 months = $4,023,734.40 a year to run SETI/BOINC !!
>
> That's alot of money, Jeez!!!!

You think so?
What about the amount below? Makes Seti look less than small change.

$149,596,046,261.00


From here

---

Grant
Darwin NT

Well it wasn´t for the money (shut down the 300 Watt heater of my waterbed and sleep in the gastroom, not telling my wife why :-) but just interest in bringing the power consumption down,.. anyway started a nice discussion !
Think i´ll buy the Arctic cooler anyway to make the processor live longer :-)

Thanx to all

---

> > Hi to all,..
> >
> > I´ve been meassuring my power consumption with and without Boinc running
> and
> > where surprised about the 40 Watt extra Boinc consumed. Is there somebody
> who
> > knows how to reduce to this ?
> > I think the 100% processor usage takes me to this high consuming
> level,...
> > Running threathmaster is no solution because i want Boinc to analyse FAST
> :-)
> > Would watercooling or an arctic cooler have a positive
> > effect on this issue ?
> >
> >
> >
> >
>
> Running an enhanced CPU cooler will not reduce your power consumption, but
> instead will increase it. Running the cpu at a cooler temperature won't
> change the amount of power the CPU uses, but the bigger 'cooler' will use more
> power moving the heat away from the cpu at a faster rate.
>
> Now, that said, running your CPU at a cooler temperature WILL extend its life
> and cut down on random errors (especially if you have overclocked it). One
> thing you might want to look at is if you are near the design current capacity
> of your computer's power supply on one or more voltages. Power Supplies work
> more efficiently when they are not near (or at) capacity, wasteing less
> electricity as heat. So, if your power supply *IS* near capacity, replacing
> it with a higher-capacity model will save you a small fraction on your power
> usage.
>
>
>
No no,..
Power supply is 450 Watt and running my PC normal (with Boinc) uses 285 Watt and without Boinc about 235 Watt. So i´ve got some spare Watts for other apps :-)

---

> > 17 'puters.... 40 watts each....
> >
> > I'M BLIND! I'M BLIND
> >
>
> Doing science for SETI and heating your home :D
>
>
>

Exactly thats what I use my PC for, it actually heats my room at night !! Isn´t that neat ? :)

---

> Well it wasn´t for the money (shut down the 300 Watt heater of my waterbed and
> sleep in the gastroom, not telling my wife why :-) but just interest in
> bringing the power consumption down,.. anyway started a nice discussion !
> Think i´ll buy the Arctic cooler anyway to make the processor live longer :-)

Hmmmm…

Why not bye watercooling for the CPU and use it to heat the waterbed?

How many watercoold CPU’s would you need to heat the waterbed?

Now the remaining problem to solve is to get rid of the noise from the power supply.

LOL :-))

---

> OK, I went and did it. I placed my multimeter in series with the cord and as
> expected the "Clamp On" ammeter wasn't very accurate at that low of a current
> draw.
>
> This puter drew from 830-912 mA (milliamps) from Bios load to the completion
> of windows loading. I shut down Boinc and it dropped to between 630-655 mA.
> This represents a delta (difference) of 228.5 mA (avg)or 27.83 watts.
>
> I'll start boinc. Yep it again went back up to 884mA. It is consistent.
>

How much does the load decrease when you turn the screen off (I'm assuming you have a LCD screen)?

C

---

Join Team MacNN

> Why not bye watercooling for the CPU and use it to heat the waterbed?
> How many watercoold CPU’s would you need to heat the waterbed?


LOL! I love this idea...use the waterbed as a giant radiator for the water cooler! As far has how many, that's pretty simple really. We know that we get at least 40 watts of heat or so with SETI running; at idle, it's probably another 20 or so. So that's 60 watts or so of heat being transferred to the waterbed; I'd bet the waterbed heater isn't much more than that. Ultimately it would depend on how well the waterbed is insulated (i.e. how big a comforter is on there). If you take the comforter off during the day, you could probably run 2-3 PC's using the waterbed as the radiator!

---

>
> > Why not bye watercooling for the CPU and use it to heat the waterbed?
> > How many watercoold CPU’s would you need to heat the waterbed?
>
>
> LOL! I love this idea...use the waterbed as a giant radiator for the water
> cooler! As far has how many, that's pretty simple really. We know that we
> get at least 40 watts of heat or so with SETI running; at idle, it's probably
> another 20 or so. So that's 60 watts or so of heat being transferred to the
> waterbed; I'd bet the waterbed heater isn't much more than that. Ultimately it
> would depend on how well the waterbed is insulated (i.e. how big a comforter
> is on there). If you take the comforter off during the day, you could probably
> run 2-3 PC's using the waterbed as the radiator!

Good guesses. Waterbed heaters are generally 100-300W from what I can tell. Assuming efficient dissipation into the "bed reservoir" itself, three or four PC's should do it. Only problem is most waterbed heaters are temperature regulating. This ability would be lost, so the actual dissipation would have to be balanced not to overheat or underheat the bed.

Actually, this idea would solve a problem that led to my worst disaster with a waterbed. I didn't fill the mattress high enough, and in the night I slept in such a way as to press a spot on the mattress down so that top, bottom, liner, and heater vinyl were all making contact. Without water dissipating the heat, the vinyl melted, welded, and ripped, including the lining, and the bed leaked many gallons before I woke up and quickly drained the rest of the bed. This system, heating the water itself, would eliminate the worry about welding the vinyl to the heater.

By the way, heat dissipation in a microchip is generally proportional to the frequency the circuit is running at and how much of the chip is actually running. Any overclocker has seen this since higher speeds need better coolers. Any digitally-run transistor (one that is expected to be running either fully on or fully off at any given time) dissipates the most heat during the time when it's changing from one state to another, hence, assuming transition times for the signal are equal across all frequencies, the higher the frequency a transistor is switched at, the more time it spends in this "resistance region" dissipating energy.

The higher power dissipation for running SETI therefore is probably due to the computer running instructions that cause more of the transistor field on the processor to change states (floating point subprocessor?) than it would otherwise running basically integer and "no operation" instructions 99% of the time.

It shouldn't be that hard to regulate the temperature of the bed, either. Find a thermocouple or other heat sensing device in the temperature range of a standard waterbed. Design a circuit / thermostat to control when to pump the heated water into the bed. The issue here now becomes, what to do with the heated water when the thermostat is saying the watertemp of the bed is too high.

Another option is to use fewer CPUs to heat the bed without the above mentioned bohemoth and use the waterbed heater to top off the heat, thereby still retaining the temperature regulation of the standard waterbed heater. This will keep the waterbed heater from running as much but still retain the water resoviour (sp? i'm sorry, i'm tired) to cool the CPUs.

Now, the question becomes, how to pipe this all into the house without making a total mess. If the PC's are in the bedroom it isn't near as big of a problem. but, for us with PC's in other rooms the plumbing aspect becomes a nightmare really quick.

> Good guesses. Waterbed heaters are generally 100-300W from what I can tell.
> Assuming efficient dissipation into the "bed reservoir" itself, three or four
> PC's should do it. Only problem is most waterbed heaters are temperature
> regulating. This ability would be lost, so the actual dissipation would have
> to be balanced not to overheat or underheat the bed.
>

---

> It shouldn't be that hard to regulate the temperature of the bed, either.
> Find a thermocouple or other heat sensing device in the temperature range of a
> standard waterbed. Design a circuit / thermostat to control when to pump the
> heated water into the bed. The issue here now becomes, what to do with the
> heated water when the thermostat is saying the watertemp of the bed is too
> high.
>
> Another option is to use fewer CPUs to heat the bed without the above
> mentioned bohemoth and use the waterbed heater to top off the heat, thereby
> still retaining the temperature regulation of the standard waterbed heater.
> This will keep the waterbed heater from running as much but still retain the
> water resoviour (sp? i'm sorry, i'm tired) to cool the CPUs.
>
> Now, the question becomes, how to pipe this all into the house without making
> a total mess. If the PC's are in the bedroom it isn't near as big of a
> problem. but, for us with PC's in other rooms the plumbing aspect becomes a
> nightmare really quick.
>
> > Good guesses. Waterbed heaters are generally 100-300W from what I can
> tell.
> > Assuming efficient dissipation into the "bed reservoir" itself, three or
> four
> > PC's should do it. Only problem is most waterbed heaters are temperature
> > regulating. This ability would be lost, so the actual dissipation would
> have
> > to be balanced not to overheat or underheat the bed.
> >
>--------------
this is the kind of thinking we need more of...ways to turn a disadvantage into an advantage...solving multiple problems with one solution...albeit a solution rube goldberg would be proud to have authored!

---

PROUD TO BE TFFE!

> It shouldn't be that hard to regulate the temperature of the bed, either.
> Find a thermocouple or other heat sensing device in the temperature range of a
> standard waterbed. Design a circuit / thermostat to control when to pump the
> heated water into the bed. The issue here now becomes, what to do with the
> heated water when the thermostat is saying the watertemp of the bed is too
> high.

If there is an original heater, the regulator could switch in secondary cooling fans by throwing a relay instead of being wired to heating coils. Really depends on the specific design of the regulator.

> Another option is to use fewer CPUs to heat the bed without the above
> mentioned bohemoth and use the waterbed heater to top off the heat, thereby
> still retaining the temperature regulation of the standard waterbed heater.
> This will keep the waterbed heater from running as much but still retain the
> water resoviour (sp? i'm sorry, i'm tired) to cool the CPUs.

Even better.

> Now, the question becomes, how to pipe this all into the house without making
> a total mess. If the PC's are in the bedroom it isn't near as big of a
> problem. but, for us with PC's in other rooms the plumbing aspect becomes a
> nightmare really quick.

Mount the PCs in the pedestal unit, sealed in plastic just in case of a leak. In our solution, there's no worry about occluding fans and such, and all the construction underneath will dampen the noise the pump we're using makes.

[EDIT] If control is really needed in another room, I've successfully attached VGA monitors with 25' extensions on them with no visible degradation of picture, and repeating amps are available. I'm wondering if an IR mouse and keyboard can be repeated by those IR to RF remote control repeaters that Radio Shack sells.

"SETI/BOINC has 159,672 Active hosts at $2.10 a month times 159,672
hosts = $335,311.20 a month to operate the entire SETI/BOINC program."

Of those 159,672 user, how many do you suppose are running multiple computers? I have 8 with two having two cpu's each for a total of 10 cpu's. I would bet that the total comuters being used is a fair amount above the 159,672.

---

Nah nah nah...theres a problem with all of this.

With all of your concern about whether the bed was radiating enough heat, and trying to keep the bed cool, you would avoid other uses for the bed itself =P

I've heard of video games creating this problem but never BOINC...

Hmmm...maybe its an alien plot!!

> "SETI/BOINC has 159,672 Active hosts at $2.10 a month times 159,672
> hosts = $335,311.20 a month to operate the entire SETI/BOINC program."
>
> Of those 159,672 user, how many do you suppose are running multiple computers?
> I have 8 with two having two cpu's each for a total of 10 cpu's. I would bet
> that the total comuters being used is a fair amount above the 159,672.
>

I'd say the number of truly active hosts is less than that.
For instance, we've got 2 hosts/computers listed but only
1 is actively crunching. The other was retired in July
but is still there because we haven't been able to delete it
due to 1 WU that's still listed for it. There's probably
quite a few users that have never crunched yet and some users
who don't even have a machine listed and some that have quit
because this "Find The Aliens Game" was no fun when they
couldn't find the start button or the joystick controls.

Dominique

---

> > I have 8 with two having two cpu's each for a total of 10 cpu's. I would
> bet
> > that the total comuters being used is a fair amount above the 159,672.
> >
>
> I'd say the number of truly active hosts is less than that.
> For instance, we've got 2 hosts/computers listed but only
> 1 is actively crunching. The other was retired in July

i'd have to vote on the side of the number being higher, i have 10 PCs currently crunching, one HT. of those, only one is 8hrs/day, the rest are 24/7.
i'd bet that there are enough ppl with these setups (i still recall the guy with 8 physical cpus in one of his boxes) that we make up for the single non-24hr and inactive crunchers.
of course, the only way to put this to rest is one of the nice developers to actually check the number of active hosts ;)

happy crunching!

---

S@h Berkeley's Staff Friends Club

> of course, the only way to put this to rest is one of the nice developers to
> actually check the number of active hosts ;)
>

You can track it yourself, the stats files are generated daily including one for all the hosts. From the bottom of the Statistics and Leaderboard page:

These sites use XML-format data exported by SETI@home.
The format is described here, and the files are here.

Download host.gz every couple of days and use it to keep track of all the hosts. Active hosts are the ones with credit that keeps changing. Beware though, even though its a respectable 19MB gzipped, uncompressed its over 145MB :)