Quote:
Originally Posted by pHaestus
You are like a bipolar person jaydee, but you flip-flop from the "good enough" to "get it right" crowd instead. Why monitor temperatures? Cooling performance is evaluated by measuring temperatures and properly analyzing their meaning. Why to bother?
|
Bipolar, that is a considerable mental problem? Anyway not nearly as bad as Bill changing from die sims to CPU test benches pretty rapidly. What is right though? No one here is capable of such obviously. We have to settle for good enough. The last 6 years here has more than proven that as we are no were closer to the answer of what is "getting it right". From your level of testing and the standard run of the mill testing the order of performance is still pretty much the same.
Quote:
Ok let's assume (here's a big one) that commercial water cooling systems are engineered and optimized to perform as well as possible at a given price point given typical space/mounting limitations. In other words that the commercial mfgrs know what they are doing. OK operate on that assumption and decide you want to build a DIY cooling loop to save money and have more fun.
|
I am going to assume you mean DIY as combining commercial products to make a loop as opposed to real DIY loop in constructing the entire loop yourself from parts you built. Anyway the tests I described will easily tell how well each work. If computer performance is degraded in any way then you have your answer.
Quote:
Given say half the budget of the commercial kit, would you want that DIY loop to perform 50%, 75%, or 90% as well as that totally optimized commercial system? The way you improve performance is by matching components, and one cannot do that without reliable test results. That is the value of properly conducted testing in my eyes.
|
But is that 50%, 75% or 90% actually benifiting the computer or your performance goals such as noise ect...? If not the question is meaningless. As said with the tests based on computer performance you will come up with IMO better answers. Instead of results based of temps they are based on actual computer performance. If the coolng system has a 75% better cooling capacity than another but the computer performance is unchanged then who cares? By the time CPU's get hot enough to were that 75% might be benificial you will have to buy a new water block at least anyway because the mounting system will probably change.
gmat mentioned cars. Lets use that for example. How are cars tested for the average end user? They are run through performance tests on a track (like overclocking and using CPU burn on a comp) and in actual driving conditions (like playing games or running vid editiong software on a comp).
How can you base the performance of a cooling kit over the performance of the complete computer system? Is the cooling system not just a part of the computer system? Why should we base performance on the capability of the cooling system as opposed to how well the cooling system actual benifits the computer?
Back to cars... After market parts are not rated on what the part is capable of but what it is capable of on the car it is designed for. Only way to find out how well that part really works is to use it on the car and see what performance increase it gives the car.
You can run all the tests you want on that part alone but in the end its real value comes with what performance gain it adds to the product it is attached to and that is the information most of the readers of reviews are looking for IMO.
More technical data can be useful for product designers but how many of those are here anymore? and what percentage of them are capable of finding the restults on their own?