Originally Posted by ebduncan
Originally Posted by KyadCK
You say that... But you just said everyone who owns a UD3 has a bad motherboard.
I will not sit back and do nothing as people make ignorant posts claiming that all things of the type suck because theirs does. Either make it clear that you are the one having problems, or don't make a claim.
Your entire statement was just as bad as someone who gets a single leaky H80 and says Corsair as a company sucks. You just said that about the entire UD3 line, even though there are three
different UD3 boards for the 900 chipsets, and you don't even have the best one.
Wow. Wrong wrong wrong wrong wrong. You should, based on your "experience" know so much better than to make a statement like that.
All I have to say is "LN2" and your statement goes completely out the window. The efficiency of a CPU under that kind of cooling is insane, and very easily noticed. Why, exactly, do you think they keep servers in 60F (at full load) rooms? They could just as easily function as high as 100F with everything going full tilt, and the company could save a lot of money paying for less A/C, but they do it anyway.
Efficiency. That's why.
Or, heck, lets go with an example that people in this thread would know best... AMD CPUs will require less voltage and power to function at a given clock speed if they are below around 55C than if they are at, say, 62C. The act of cooling it better allows the CPU to use less power, and thus generate less heat, and keep the CPU even cooler. This is a well known and proven fact, at minimum for the Ph II line.
No one cools their computer 24/7 with ln2 its not practical. Before you say phase change, how many people actually have it? your listing extremes which do not fit even remotely close to real world. IE 99.9999% of computers in the world are air cooled. IF your going to try and pick apart a blanket statement, treat it as such.
Why do you keep servers at a cooler temperature? fatigue less heat, less friction things last longer. Its not efficiency, its a cost perspective, they spend less on cooling, than they do with increased failure rates due to higher temperatures.
Your last example makes sense, however your adding variables. Yes if you run at a lower temperature you might be able to get away with running less voltage. However strictly speaking we are comparing cooling systems, and how it effects the ambient air. Given the fact the computer and its settings are not changed. If you give it less voltage then less heat is created, wouldn't be a valid control in an experiment .
way to make yourself look like an ass.
They don't need to. You said "Regardless of cooling". LN2 is cooling. Therefore you are wrong. Do not make blanket statements your training should have taught you not to make.
They do not do it to make the parts last longer. I get that few people understand this, but that's not how the Enterprise computer world works, and not how Enterprise chips work. They are rated for higher temps, they are binned better, they are guaranteed to work within spec, and regardless of what the company does they must
run at those higher temps. If a chip dies in the enterprise world it cost the company nothing. They make one phone call and there is a replacement in the hands of the operators the next day, and no that doesn't hurt them because N+1. Always have a backup, live, ready for use. Always have a mirror.
I don't care if I'm adding variables, try to keep up. There is more to the argument of cooling than just "it's colder on the CPU", and if you can't take those into consideration, it's a good thing I'm not blindly listening to you.
The topic was "better cooling so the room can be colder". I never said the CPU would be at the same voltage, I never said I wouldn't modify what I could. You just assumed it would, and told everyone (who is right) they are wrong.
Originally Posted by M3TAl
Not sure what's so hard to understand. Doesn't matter if we're using the stock cooler at 62C or custom water at 35C when 100% loaded the CPU is generating the same amount of heat. One just does a better job of getting that heat away from the CPU. Either way the chip is outputting the same amount of heat.
Edit: really think you guys are over-thinking it. I remember the first time I pondered this same question, was years ago in the socket 939 days. I was telling my Dad I was going to put some better cooling on the CPU so I would get better temps and room would heat up less...
You know what my Dad did? He laughed at me (petroleum engineer)! He told me it didn't matter what cooling it was, same amount of heat will still be in the room. So I had to think about that for a moment and I came to the conclusion he was correct.
You should have laughed back, as he failed to take into consideration efficiency and voltage into that scenario. He is right for Petroleum, he is not right for CPUs.
Originally Posted by d1nky
so for example if ya cancelled out some of these changing variables (efficiency/temp) and kept it fixed, the other guys would be right
and if you didnt change voltages.....
If Efficiency and voltage changes were not a thing (What I like to call the "Physicist's Perfect Room"
), then their theories of something putting out the same heat would be true. Unfortunately for them, those variables do exist, and writing them off or not taking them into consideration is a massive error. It's like shooting at a target and forgetting bullet drop/gravity exists.
Originally Posted by Vencenzo
Do you know what the ud3 rev3 and gandalf have in common?
"You may not pass!" this specific heat limit because screw you we're going to make APM standard and you must have it on to have a modified multiplier?
Originally Posted by bond32
No, he is right and you are wrong. The computer "produces" (transfers) the same amount of heat regardless of the cooling. The purpose of the cooling is to "transfer" said heat to the air. Look up Newton's laws if you don't believe it. Energy is neither created nor destroyed which means Energy in - Energy out + Energy Generated equals change of energy with respect to time. What does this mean? Means the heat is still there for better cooling solutions, but those are able to move the heat and transfer it to the ambient much quicker as opposed to an air cooled solution.
Consider a very high efficiency heat pump where even in -10 degree C weather, it STILL is able to heat the house without kicking on the backup heat. How is this possible? Because even in -10 degree C temperatures, there is STILL heat that can be pulled from the air.
I'm sorry bond, but you failed to read the last sentence entirely. You did not bring efficiency of the CPU
and possible voltage changes into the equation.
Once again, compare the efficiency of a CPU under water vs a CPU under LN2 if you want a bright as day example you can study.