IMPORTANT HEADS-UP AND REQUEST
I've followed the ads and hype about the CNPS-9500 for several weeks now.
It was not included among the ten or so coolers evaluated in September '05 Maximum PC. In that review, Max PC showed again that its reviews are a bit "loose-jointed" or lacking in explicit across-the-board evaluation results. Disregarding any judgment about cooler size, ease of installation, weight and other factors, the ultimate measure of a cooler's effectiveness is its thermal resistance -- a number which summarizes information about the difference between idle and load temperatures for a given processor's measured maximum heat-leakage. The Maximum PC article at least provides enough data to compute the thermal resistance value: the tables give idle and load temperatures for each cooler measured with a specified AMD processor and Intel processor, and the rated heat leakage in watts for either processor can be found at the respective manufacturer's web-site.
But the Zalman 9500 -- as I said -- was not included in that review. And the hype in Zalman's own ads leave much to be desired. Instead of showing the thermal resistance approaching a minimum limit as fan speed increases, the ad only shows how temperature decreases as fan speed increases, and the graph in the ad presents dotted-line extrapolations of questionable accuracy.
Zalman insists that the new heatpipe cooler does a better job than off-the-shelf water-cooling kits. Whether the cooler actually performs that well is one thing. What we need to look for are reviews of the product which benchtest the cooler and present thermal resistance values for it. Whether or not it actually trumps water-cooling, if the thermal resistance values are significantly less than those of the ThermalRight XP-120 and -90 units, then we will know for sure that the Zalman cooler is the best on the market.
The ThermalRight XP-120 shows a minimum thermal resistance of 0.167, and I was able to independently verify this to confirm two reviews at two different German web-sites last year. That beats the Zalman CNPS-7700-Cu (0.19) and the CNPS-7000-Cu (0.22).
Keep an eye out for reviews which present thermal resistance benchmark results. It will help us if the first such reviews are posted in this forum.