I don't know about the X2 3800, but I have Socket 939 4400 and 4800 and they both come with an excellent HSF (Heatsink-fan] assembly (it's a heatpipe design with a very powerful fan). I tries a few whizzbang-looking coolers to see if I could get better cooling out of them, and the stock HSF beat them every time. I'm sure there are coolers out there that are better, but you would need to go right to the high end.

The only grumble I have about the stock HSF is that it is temperature-controlled. That is, although the fan can spin up to about 5115RPM, it will only do that when the air temp in the case is up to about 45 degrees or more. By this time, if you are overclocked, your CPU is labouring along at about 63 degrees. I had to blow a hairdryer at the air intake port to get full speed out of the CPU fan.

Since I prefer to have the fan trying to run at full speed and then slow it down with speedfan or even the Q-fan option on ASUS motherboards, this isn't what I wanted. The Fan speed as it is depends on the temp of the case air rather than the temp of the CPU.

I managed to fix this somewhat by turning the fan over so it sucks air through the heatsink rather than blowing on it. This way, the air has already been used to cool the CPU and is a few degrees hotter. What's more, as the CPU heats up so does the air (somewhat). Using this technique, I managed to get a good 2 degrees cooler than with the fan the other way, and a good 5 degrees better than any of the after market HSFs that I tried.

Be aware that many/most motherboards are designed with the assumption that there is a CPU fan blowing air down through a heatsink and out across the capacitors, which also need cooling. It is not a worry for me, because the caps are sitting right next to a big 120mm case fan that is sucking a lot of air across them anyway. You should think about this though. If you reverse the CPU fan you probably reduce the rate of airflow across the capacitors if that is the only thing stirring the air.

Replacing the heat-conductive pad on the cooler with well-spread (i.e. *very* thinly spread) thermal paste seems to help a bt as well, but I have not quantified it.

Bottom line: I would not get an after-market HSF for an Athlon 64 X2 unless I found I really needed it, and then I would go for a very high-end one because the stock HSF is really pretty good.

[.....]
Stock cooling solution is guaranteed for 3 years on the boxed products. If the CPU or cooling solution stops working, you send it back to AMD and it's replaced for the cost of shipping from you to AMD... If you don't use a stock cooling solution, you're essentially voiding your warranty, but of course it's possibly not easy to prove that - but you may need to send in the heatsink/fan and if it looks like you haven't used the one supplied by AMD, they may refuse to replace it for you...

If you're going to overclock, then you may need to go to a different colling solution - and then you definitely have no right to warranty on the processor anyways... ;-) [Again, it depends on what tests are done to verify the claim here, it may not be easy to prove whether a processor was run at a higher clock-speed than it was intended for...]
[..]
--
Mats

---

--miw

I managed to fix this somewhat by turning the fan over so it sucks air through the heatsink rather than blowing on it. This way, the air has already been used to cool the CPU and is a few degrees hotter. What's more, as the CPU heats up so does the air (somewhat). Using this technique, I managed to get a good 2 degrees cooler than with the fan the other way, and a good 5 degrees better than any of the after market HSFs that I tried.

Ouch that sounds weird, but if it does what you want so be it. It seems to me your goal is to get the temps down no matter what. My goal is more to keep the temps in line even if I overclock like hell. In that regard I am very satisfied with ARCTIC COOLING Freezer 64 Pro, which is quite inexpensive as well. It's quite big though.

I managed to fix this somewhat by turning the fan over so it sucks air through the heatsink rather than blowing on it. This way, the air has already been used to cool the CPU and is a few degrees hotter. What's more, as the CPU heats up so does the air (somewhat). Using this technique, I managed to get a good 2 degrees cooler than with the fan the other way, and a good 5 degrees better than any of the after market HSFs that I tried.

Ouch that sounds weird, but if it does what you want so be it. It seems to me your goal is to get the temps down no matter what. My goal is more to keep the temps in line even if I overclock like hell. In that regard I am very satisfied with ARCTIC COOLING Freezer 64 Pro, which is quite inexpensive as well. It's quite big though.

No. My goal is to get the temp down to about 56 degrees even when the aircon is off and I am overclocking by 10%. Turning the fan over basically put the temp sensor on exit rather than on entrance, which improves the control laws of the system. seems silly to have a fan that can to over 5000RPM idling along at 3200 RPM when the CPU is cooking at 60+ degrees. When I get to a decent shop I am going to get a Scythe Ninja cooler and give that a try. Nett effect of turning the fan over was an increase from 3300RPM to about 4000RPM when the aircon is off. (and a 2-3 degree temperature drop on the CPU.)

---

--miw

No. My goal is to get the temp down to about 56 degrees even when the aircon is off and I am overclocking by 10%. Turning the fan over basically put the temp sensor on exit rather than on entrance, which improves the control laws of the system. seems silly to have a fan that can to over 5000RPM idling along at 3200 RPM when the CPU is cooking at 60+ degrees. When I get to a decent shop I am going to get a Scythe Ninja cooler and give that a try. Nett effect of turning the fan over was an increase from 3300RPM to about 4000RPM when the aircon is off. (and a 2-3 degree temperature drop on the CPU.)

Could you cut the thermal resistor out off and connect the wires together directly? I think that'd make the fan run at 100% so you could control it using speedfan or something similar on the motherboard.

---

ASUS M2N32 WS Professional nForce590 SLI
with excellent features
------------------------

---

No. My goal is to get the temp down to about 56 degrees even when the aircon is off and I am overclocking by 10%. Turning the fan over basically put the temp sensor on exit rather than on entrance, which improves the control laws of the system. seems silly to have a fan that can to over 5000RPM idling along at 3200 RPM when the CPU is cooking at 60+ degrees. When I get to a decent shop I am going to get a Scythe Ninja cooler and give that a try. Nett effect of turning the fan over was an increase from 3300RPM to about 4000RPM when the aircon is off. (and a 2-3 degree temperature drop on the CPU.)

Could you cut the thermal resistor out off and connect the wires together directly? I think that'd make the fan run at 100% so you could control it using speedfan or something similar on the motherboard.

yes. Maybe. Looked very fiddly to do and I haven't been brave enough to try as yet. :-) But "disable the temp control to 100% all the time on the fan" would be the result I am looking for.

---

--miw

The only grumble I have about the stock HSF is that it is temperature-controlled. That is, although the fan can spin up to about 5115RPM, it will only do that when the air temp in the case is up to about 45 degrees or more. By this time, if you are overclocked, your CPU is labouring along at about 63 degrees. I had to blow a hairdryer at the air intake port to get full speed out of the CPU fan.

I believe (strongly) that this is not an AMD decision, but rather a motherboard decision to put a pulse-width-modulator on the fan. This in turn is decided based on the fact that people would complain that the fan is making too much noise when run at full speed.

However, a CPU running at 63'C is fine, no worries at all. AMD processors have a max temp of 70'C, which is well above what you're seeing, and it should work fine for many years at 63'C. Trying to run the processor cooler than this is just a waste of energy (spinning the fan faster). But if you want to "fix something that isn't a problem", then go ahead.... ;-)

--
Mats

---

The only grumble I have about the stock HSF is that it is temperature-controlled. That is, although the fan can spin up to about 5115RPM, it will only do that when the air temp in the case is up to about 45 degrees or more. By this time, if you are overclocked, your CPU is labouring along at about 63 degrees. I had to blow a hairdryer at the air intake port to get full speed out of the CPU fan.

I believe (strongly) that this is not an AMD decision, but rather a motherboard decision to put a pulse-width-modulator on the fan. This in turn is decided based on the fact that people would complain that the fan is making too much noise when run at full speed.

However, a CPU running at 63'C is fine, no worries at all. AMD processors have a max temp of 70'C, which is well above what you're seeing, and it should work fine for many years at 63'C. Trying to run the processor cooler than this is just a waste of energy (spinning the fan faster). But if you want to "fix something that isn't a problem", then go ahead.... ;-)

--
Mats

I just realized that I looked at the AM2 model, which has a higher max temp than the 939 model - but it's still fine to run it up to the max temp, whether it's overclocked or not [of course, overclocking MAY not work at max temp, but that's a different story].

--
Mats

---

However, a CPU running at 63'C is fine, no worries at all. AMD processors have a max temp of 70'C, which is well above what you're seeing, and it should work fine for many years at 63'C. Trying to run the processor cooler than this is just a waste of energy (spinning the fan faster). But if you want to "fix something that isn't a problem", then go ahead.... ;-)

--
Mats

Does anyone know what difference the temperature of a CPU makes to its power consumption? I remember reading that the water-cooling system for the xbox-360 reduced the overall power consumption, even though there was the addition of a water pump...

---

I believe it's a direct relactionship between heat and power consumption. I can't think of any other way for the power to be dissipated other than thermal.

---

Does anyone know what difference the temperature of a CPU makes to its power consumption? I remember reading that the water-cooling system for the xbox-360 reduced the overall power consumption, even though there was the addition of a water pump...

The thermal loss of a transistor depends on the temperature, this sure applies to integrated circuits like CPUs as well or even more. So if you are able to keep the CPU cool, the CPU will produce a little less heat by itself, making the cooling even more efficient.

I do not know how much difference in the power consumption it makes, to have 10C less, but it might be higher than the difference between the consumption of fan and water pump, thus reducing the overall consumption.

p.s.: I found some formulas on the german Wikipedia, 11 kelvin more causes a bipolar transistor to double the collector current (sounds quite much to me but as I do not know better, I have to believe it), CPUs do not have this type of transistor, FET / MOS have different characteristics but basically the behaviour must be similar

Does anyone know what difference the temperature of a CPU makes to its power consumption? I remember reading that the water-cooling system for the xbox-360 reduced the overall power consumption, even though there was the addition of a water pump...

http://www.engadget.com/tag/overheating/
The mention of the cooler running Xbox 360 is for a 65 nm cpu version; which can use a lower voltage than the cpu using the 90nm process. The paragraph mentions some of the water cooled Xbox 360s out there. I visited a few of the Xbox 360 water cooling pages that described having a system that was 50 degrees Fahrenheit cooler than stock after an hour playing a particular game. They didn't mention anything about electrical usage, though. Did I miss the writeup that mentioned electrical usage?

---

currently formatting partition 1. Can't hardly wait. Atleast I've seen no smoke so far the CPU cooling device/fan is rather impressive with heatpipes and everything. Very quiet (when not under load, but I'll fix that soon), I can definitely hear the fans of the other puters in this office over this one, even with the side door off.

bump

bump

Why?

---

Nemesis n. A righteous infliction of retribution manifested by an appropriate agent.

Seems pretty clear to me that Casey is changing the conversation de jour. Back to things that actually will help CRUNCHING Rosetta... on the Crunching forum... novel concept.

---

Add this signature to your EMail:
Running Microsoft's "System Idle Process" will never help cure cancer, AIDS nor Alzheimer's. But running Rosetta@home just might!
https://boinc.bakerlab.org/rosetta/

Seems pretty clear to me that Casey is changing the conversation de jour. Back to things that actually will help CRUNCHING Rosetta... on the Crunching forum... novel concept.

Indeed :)


The only problem is that the "bumping up" trick of trying to bring down to page 2 status other threads can be played by all . If abused, it will become as useless and insignificantly annoying as the + and - buttons have become in this MB.