Please move distributed computing discussions to comp.distributed

Ed wrote:
"Stratcat" wrote in
:


The above post may be an attempt to build trafffic on an otherwise
low traffic group. A Google search shows 651 threads and 2020 posts
[...]
SAH was definitely chartered as on-topic at the time of alt.sci.seti's
[...]
Maybe the real question is do some of the posts here really need to be
cross posted? While they have to do with seti, the isues with boinc may
also be distributed computing related.


It does get rather irritating for almost all posts to be cross posted!

Keep general and s@h stuff on alt.sci.seti and more general science and
refs/links/astro to sci.astro.seti...?


IE: DO NOT CROSS POST!

I would guess most people subscribe to both groups just as I do.

Regards,
Martin

--
---------- OS? What's that?!
- Martin - To most people, "Operating System" is unknown & strange.
- 53N 1W - Mandrake 10.0.1 GNU Linux
---------- http://www.mandrakelinux.com/en-gb/concept.php3

Apollo wrote:
~misfit~ wrote:

I agree about todays HOT cpu's consumption, but as Matt was talking
about a 200W psu I was thinking PIII / Low-end athlon/duron systems.
I have no experience of recent Intel procs, but I don't think I'd put
a PIV system together using a 200W psu ;o)

And yet some builders do, mainly outfits like Compaq. I don't know about
latest systems but I've seen older P4's with 200 w PSUs.

It all depends on your CPU. A PrescHOT has a thermal dissapation of
about 103 watts so it's fair to assume that it's using at least that
amount plus a little more. Motherboard circuitry uses a fair few
watts (have you ever put your finger on the power FETs on a running
PC, or wondered why northbridges have heatsinks/fans?) as well. An
AMD XP Athlon 32-bit dissapates around 83 watts.

I'm curious now, any idea what my cpu would consume; mobile XP2400
35W @
1.35v standard vcore - set at 1.93v. There must be some correlation,
it can't be linear or it would only be around 50W and I wouldn't need
my wetware setup.

Yeah definitely not linear. I can only speculate but I'd say it would be
around/above the 100 watt mark, possibly as high as 140 watt.

I wish I knew someone with one of those sockets that measure power
consumption, I'd definately borrow one for a while just to see what
I'm using, I wonder if any tool hire shops have them?

I've thought the same thing. I'd like to know *exactly* what a PC uses. I've
been involved in discussions like this in several NG's over the years and
nobody seems to have a definitive answer. Hard data.

Cheers,
--
~misfit~

Apollo wrote:
~misfit~ wrote:

I agree about todays HOT cpu's consumption, but as Matt was talking
about a 200W psu I was thinking PIII / Low-end athlon/duron systems.
I have no experience of recent Intel procs, but I don't think I'd put
a PIV system together using a 200W psu ;o)

And yet some builders do, mainly outfits like Compaq. I don't know about
latest systems but I've seen older P4's with 200 w PSUs.

It all depends on your CPU. A PrescHOT has a thermal dissapation of
about 103 watts so it's fair to assume that it's using at least that
amount plus a little more. Motherboard circuitry uses a fair few
watts (have you ever put your finger on the power FETs on a running
PC, or wondered why northbridges have heatsinks/fans?) as well. An
AMD XP Athlon 32-bit dissapates around 83 watts.

I'm curious now, any idea what my cpu would consume; mobile XP2400
35W @
1.35v standard vcore - set at 1.93v. There must be some correlation,
it can't be linear or it would only be around 50W and I wouldn't need
my wetware setup.

Yeah definitely not linear. I can only speculate but I'd say it would be
around/above the 100 watt mark, possibly as high as 140 watt.

I wish I knew someone with one of those sockets that measure power
consumption, I'd definately borrow one for a while just to see what
I'm using, I wonder if any tool hire shops have them?

I've thought the same thing. I'd like to know *exactly* what a PC uses. I've
been involved in discussions like this in several NG's over the years and
nobody seems to have a definitive answer. Hard data.

Cheers,
--
~misfit~

"Matt Giwer" wrote in message
.. .

snip

At 24/7 you are not temperature cycling your computer which should
extend its life.

Not sure what your implying here. Tempaerature cycling is exactly what
will SHORTEN an electronics device's life. It's better to run electronics
in a steady state environment, within thier specs.

Electro-mechanical devices may last longer due to thier mechanical
nature, but in general, a great deal of stress occurs during the temperature
cycling phases, and the in-rush of currents that occur when a device
powers up.

Unless there's something you're aware of, that I'm not.
--
Strat

"Matt Giwer" wrote in message
.. .

snip

At 24/7 you are not temperature cycling your computer which should
extend its life.

Not sure what your implying here. Tempaerature cycling is exactly what
will SHORTEN an electronics device's life. It's better to run electronics
in a steady state environment, within thier specs.

Electro-mechanical devices may last longer due to thier mechanical
nature, but in general, a great deal of stress occurs during the temperature
cycling phases, and the in-rush of currents that occur when a device
powers up.

Unless there's something you're aware of, that I'm not.
--
Strat

"Apollo" wrote in message
...
Matt Giwer wrote:

As to cost, you cannot be drawing more than the wattage of your
computers power supply, probably closer to half unless you have four
hard drives and peripherals drawing power.

200 Watts X 24 hours = 4.8 KW-HR X cost per kw-hr

Most new psu's are 75% efficient at best, older ones even worse, so a 200W
psu, under full load, draws about 250W. I do agree that consumption at
full
load with only 1 hd and a low power gpu is probably around half or just
over
half of the stated consumption.

Agreed. Your power numbers are for an 80% efficiency factor, BTW.
--
Strat

"Apollo" wrote in message
...
Matt Giwer wrote:

As to cost, you cannot be drawing more than the wattage of your
computers power supply, probably closer to half unless you have four
hard drives and peripherals drawing power.

200 Watts X 24 hours = 4.8 KW-HR X cost per kw-hr

Most new psu's are 75% efficient at best, older ones even worse, so a 200W
psu, under full load, draws about 250W. I do agree that consumption at
full
load with only 1 hd and a low power gpu is probably around half or just
over
half of the stated consumption.

Agreed. Your power numbers are for an 80% efficiency factor, BTW.
--
Strat

"Stratcat" wrote in
m:

Not sure what your implying here. Tempaerature cycling is exactly what
will SHORTEN an electronics device's life. It's better to run
electronics in a steady state environment, within thier specs.

Electro-mechanical devices may last longer due to thier mechanical
nature, but in general, a great deal of stress occurs during the
temperature cycling phases, and the in-rush of currents that occur
when a device powers up.

Yep. There is a fire house, maybe in NJ, that had a light bulb that
Edison built still burning into at least the 90's. Don't know if it is
still working, but they never turned it off.

The trade off with PCs has always been power usage vs. the life of SOME
parts. If you keep a disk drive running all of the time, it will
probably fail sooner then if you powered the machine off. Bottom line is
that the failure issues for computers are not as simple as many people
think.

--
Ed

http://www.geeks.org/~ed/Usenet_Servers.html
strip to reply

"Stratcat" wrote in
m:

Not sure what your implying here. Tempaerature cycling is exactly what
will SHORTEN an electronics device's life. It's better to run
electronics in a steady state environment, within thier specs.

Electro-mechanical devices may last longer due to thier mechanical
nature, but in general, a great deal of stress occurs during the
temperature cycling phases, and the in-rush of currents that occur
when a device powers up.

Yep. There is a fire house, maybe in NJ, that had a light bulb that
Edison built still burning into at least the 90's. Don't know if it is
still working, but they never turned it off.

The trade off with PCs has always been power usage vs. the life of SOME
parts. If you keep a disk drive running all of the time, it will
probably fail sooner then if you powered the machine off. Bottom line is
that the failure issues for computers are not as simple as many people
think.

--
Ed

http://www.geeks.org/~ed/Usenet_Servers.html
strip to reply

Raj Rijhwani wrote:
On Monday, in article

"~misfit~" wrote:


Raj Rijhwani wrote:

On Sat, 31 Jul 2004 11:49:07 +1200, in article

"~misfit~" wrote:


Yes. However most interest in "SETI" in newsgroups is about the DC client.


Thus swamping any other discussion that the prior participants in the
newsgroup set it up to discuss. If one particular project comes to
dominate the discussion, any new initiatives are going to be stifled.
[...]


BOINC is the most significant thing happening at the moment. A lot of
posts on the subject is to be expected.


There is a question of categorisation... Should a DC project be
catagorised by the _purpose_ of the project, or by the DC architechture
supporting it?

I'm sorry, but there's likely a lot more interest in whatever project as
a whole rather than the specific computer science supporting it.

I'll guess that you'll find BOINC comments and hints and tips spread
across the forums of many projects that commonly use BOINC support.


And just one post recently in c.distributed...

Any new 'grid' computing ideas anyone for over on c.d.?

Regards,
Martin

--
---------- OS? What's that?!
- Martin - To most people, "Operating System" is unknown & strange.
- 53N 1W - Mandrake 10.0.1 GNU Linux
---------- http://www.mandrakelinux.com/en-gb/concept.php3