I'm still not convinced. Have a look at the string of reviews at Download.com:

http://www.download.com/SETI-home/3640-2400_4-10488045.html

Are all of these folks in denial, too?

Or has SETI@home evolved from the friendly, make use of your unused cycles app, to a cpu-cooking monster?

Ask yourself honestly, "Am I constantly upgrading my system to keep up with SETI@home demands?"

And this, from Hardwareanalysis.com:

http://www.hardwareanalysis.com/content/article/1770/

"We measured temperatures in a typical, well ventilated, mid-tower case with the processor running at full tilt crunching away at Prime95 and SETI@home after 30-minutes of settling in. To give you an indication we’d also like to note that idle temperatures where typically 10 to 15 degrees lower than full load temperatures."

In all fairness, I have observed some BOINC applications causing the CPU to run hotter than others, e.g. CPDN runs about 4 degrees C hotter than SETI or Einstein, which in turn run about 2-3C hotter than Predictor.

I have a similar situation with BOINC Simap, I was just about to open a thread to see if anyone knew what´s the cause of this.

Apparently some applications use more FP math than others, causing that area to become hotter & the overall CPU temp to become slighty hotter. (eg batch of faulty Opterons).

I guess I no longer have to open it ;):b:

Here´s a screenshot where you can clearly see the difference in t° with the other Boinc Projects. SETI, Einstein, World CG, LHC, don´t cause this or the difference is minimal in this scale.

CPU: Celeron D 2.66@3.01Ghz with stock cooling.

---

Hi all

I just copy you the temperatures obtained with Everest in my PC

--------[ EVEREST Home Edition (c) 2003-2005 Lavalys, Inc. ]-------------------

Versión EVEREST v2.01.347/es
Sitio Web http://www.lavalys.com/
Tipo de informe Informe rápido
Ordenador RICARDO
Generador ECOM
Sistema operativo Microsoft Windows XP Home Edition 5.1.2600 (WinXP Retail)
Fecha 2006-05-16
Hora 14:17


--------[ Sensor ]-------------------------------------------------------

Propiedades del sensor:
Tipo de sensor Winbond W83627HF (ISA 290h)

Temperaturas:
Placa base 33 °C (91 °F)
Procesador 56 °C (133 °F)
Aux 22 °C (72 °F)
WDC WD800BB-00JHA0 35 °C (95 °F)

Ventiladores:
Procesador 7500 RPM

Valores de voltaje:
Núcleo CPU 1.26 V
+3.3 V 3.39 V
+5 V 5.21 V
+12 V 12.16 V
-12 V 1.63 V
-5 V 2.59 V
Puesta en espera +5 V 5.02 V
Batería VBAT 3.25 V

Also see the report of my processor

Placa base:
Tipo de procesador Intel Celeron D 330, 2666 MHz (20 x 133)
Nombre de la Placa Base Desconocido
Chipset de la Placa Base Intel Brookdale-G i845GEV
Memoria del Sistema 247 MB (PC2700 DDR SDRAM)
Tipo de BIOS AMI (09/07/04)
Puerto de comunicación Puerto de comunicaciones (COM1)
Puerto de comunicación Puerto de impresora ECP (LPT1)

Hope this help you with the temperatures.

P.S.: above temperatures running S@H enhanced

Regards to all
Ricardo

---

I'm still not convinced.

The CPU of your computer always runs at 100%. It can only do that or 0% when it has been shut down.

Yet at times it isn't doing anything, aka when it isn't under load, it still runs at 100% yet it's "idling". Microsoft wrote a special program for that, called the System Idle Process to show you those cycles in Windows Task Manager. The CPU still runs at 100%, but it's not under load, so it isn't doing millions of calculations per second.

Then you go run Seti, which is just a mathematical program that tells your CPU to do those millions of calculations per second. And so your CPU starts to heat up from doing that.

At such a point it is always best to make sure your CPU fan is clean, that your CPU heatsink is clean and that the CPU has a good contact with the heatsink. The heatsink will take over the heat from your CPU and tries to disperse this into your case, where hopefully fans will take this hot air away and cool air from the bottom of your case can cool the heatsink again. The CPU fan transports the heat away at a greater spead, by sucking the hot air from the top of the heatsink away.

Any dust clogging up the fan or heatsink will make sure your CPU gets very hot.
Not enough fans inside your case make sure that heat lingers inside your case and everything will heat up. PSU fans that are clogged up won't suck the hot air out of your case either.

So make sure your fans are clean and that they aren't blocked by cables, which hamper airflows inside your computer.

Try playing a 3D intensive game with BOINC off. See how much that will heat up your CPU. Same thing.

---

Review put on: Jan 05, 2005, 08:30 AM

So either he was still running Seti Classic at the time, or just plain normal Seti 4.18

Yes, Seti will heat up your CPU. Anything that doesn't IDLE will heat up your CPU. All you need to make sure of is that your CPU can get rid of that heat. That's not bullshit, but very simple.

If you don't want to heat up your CPU, turn off your computer and store it in a cool cellar.

---

Yes, Seti will heat up your CPU. Anything that doesn't IDLE will heat up your CPU. All you need to make sure of is that your CPU can get rid of that heat. That's not bullshit, but very simple.

I'm big enough to admit when I'm wrong: as it turns out, when I removed the cpu fan, I found the copper cooling fins of my cpu almost completely caked with dust.

So, from running at 83 degrees C, to running at 73 degrees C running Prime95 or SETI@home or Einstein@home throttled down to 10% of cpu apiece (with the excellent ThreadMaster GUI), I'm now running SETI@home at 56 degrees C using 99% of cpu.

You can bet your bottom dollar, though, that I'm going to use ThreadMaster to find a more optimal cpu usage percentage than the default of 99 (or what many mistakenly call 100).

Kindly reserve judgment till you have your own experimental results using ThreadMaster or some other kind of cpu throttling to find (or not) better cpu usage-to-WU output ratio.

... when I removed the cpu fan, I found the copper cooling fins of my cpu almost completely caked with dust.

That'll do it.
No air flow, not heat transfer, no cooling.

Kindly reserve judgment till you have your own experimental results using ThreadMaster or some other kind of cpu throttling to find (or not) better cpu usage-to-WU output ratio.

?
The less time BOINC has the CPU for, the less time it has to crunch. The less time it has to crunch, the lower your WU throughput (and hence credits per hour).

BTW- Threadmaster doesn't throttle the CPU, it throttles programmes (well, threads).

---

Grant
Darwin NT

... when I removed the cpu fan, I found the copper cooling fins of my cpu almost completely caked with dust.

That'll do it.
No air flow, not heat transfer, no cooling.

Aye, Grant, my friend, one more hard-headed, arrogant Yank a bit bloodied but better for it, and less ignorant!