[bobo5195]

I’ve been noticing around the web and the world in general decline in buzz for computers. Forums seem to be getting smaller, less stuff seems to be coming out. There is a growing band of people certainly at the top end who know a lot of things, that PC’s are generally fast enough and they are not improving very fast these days as well. To be honest short of a few intensive tasks like games, CAD/CAM, sciency stuff, media encoding etc there is no need for anything faster than a p3 1ghz. Prices for something a little bit faster have gone up exponentially (I’m thinking SLI here). This was always the case to a certain extent but you can now see more engineering effort into getting the best, dual video cards, quad channel ram etc. As Bob Colwell pointed out in his presentation there is not really much left in faster CPU designs all the easy stuff has been done. Process scaling is hitting a little wall and heat has become a major issue, so things aren’t going to get faster. With out exponential increases in speed there is very little much more you can do with computers.

In cooling, coolers have got a lot bigger and a lot better. Looking at a modern thermalright cooler and I can’t see much of a way to do it better. In essence it is very close to a water-cooling setup, can dissipate a shed load of heat and can do it quietly. Waterblocks have similarly plateaued a little it would be hard to get anything much better there. Rads maybe there is a little optimising to do but things are good. As such pure performance has given way to aesthetics, bling(which I count as bad design but looks flash) cost/efficiency trade offs and silencing. As people have lost faith with computers time and effort which someone is willing to put in to make their computer fast and a little difference has fallen.

However there is a big but to all this dead endness. People like things silent and it is extremely hard to make a top end dual core, dual video card silent on air. Also water cooling is getting quite refined and cheaper. It is beginning to optimise for cost as swiftechs apogee shows. This could well be the best time for water-cooling there is ever going to be. You have ultra high power densities in small computer cases with low noise as a requirement. This is the stuff that water-cooling is good at. Watercooling should make case design better as you have to worry less about heat so can do something a little difference. There is currently a big hole in the market for a case that passively water cooling rad in itself. This kind of thing could be made cheaply and would completely kill a restorator as it does not need a giant phallus by your computer case.

Similarly if CPU’s have plateued then OCing becomes more attractive. A OCed A64 from a year odd back can still hold its own against the fastest cpu today. A good solution has already been out on the market. TEC’s with a temperature cut off to prevent below dew point cooling (condensation). This allows maybe an extra 100Mhz with only a little extra effort. More importantly it means that your waterblock design can be cheap because the TEC helps with c/w numbers.

As for what we experts or at least slightly better educated geeks can do on this forum there is in fact quite a lot. Less going on is a good thing for in-depth analysis. The systemcooling review of the apogee is a prime example it took time and covers everything. There is no way that could happen if there was a fast moving market and such. As things get more complicated knowledge becomes more important and the details are magnified. The apogee case is an example of that (and I hope this is not going to start a flame war by saying this). The apogee does perform on the same level as a storm with a normal IHSed CPU, not because it’s a very good cooler but because the I H S numbs performance so it is good enough. To find this out you need to be smarter than the average bear. As a forum there are a number of things that still need to be done before we can write the book and lay everything to rest. Firstly a good flow calculator (LHG v2 approximator) needs to come out for one, that allows DIYers to compete with kit solutions effectively. If a person at home can optimise then why get a kit. I enter what I need, a few clicks and bang cheapest solution for the price comes up. Should be able to beat a kit every time. There are still some tricks to be done on waterblocks and the overall design of watercooling kit so I would not worry about things going dead.

End miss-spelled crap grammar rant