Noctua NH-D14 CPU Cooler
Performance |
 My performance rating is going to be different out of necessity. I don’t have anything that could compete with this cooler. So, I’m going to apply a real world, tangible benefit for people to consider while looking at these results. I recently built a system based on an Intel i7-975EE CPU. It is a heat HOSS. I started off using a Thermaltake heatsink. That worked ok… until I started using my machine for a SMP client Folding@Home processor. And it rocked! *Until* I would saturate the heatsink to the point where it could not exhaust the heat from the CPU fast enough. Eventually, the cores would hit a temperature that forced the machine into a thermal shutdown. So, I would loose hours of potential folding time depending upon when this shutdown happened. Sometimes it happened in the wee hours of the morning, and if I did not check my machine before going to work, it could be powered off 12-16 hours until I got back to it the following evening. This annoyed me from a productivity standpoint, but it also concerned me in that I was letting my CPU get too hot, and that was shortening the life span of the silicon. Not something I want to do with that particular high end tech. So, I upgraded to a Xigmatech S1283 “Dark Knight” heat sink. This bought me time, but I would still have thermal shut downs. Below is the functional graphs and temperature info that I captured while using each heatsink.  Notice that the bottom fell out of the chart when comparing the minimum “idle” values. When the CPU kicked it up into the “fully utilized” range though, the Noctua “Max” temps beat the Dark Night “Minimum” measurements!!! In other words, this is a very efficient heatsink! Here is a graph designating weekly productivity. My PPD/PPW totals shot up once the SMP Folding was properly tweaked and getting the right work units. However, you can also see drops in my graphs that are representative of production problems (machine powering off, thermal protection).  You can see through the month of March that I had lots of difficulty, but getting into April, the PPD production starts skyrocketing! The machine became a lot more reliable and was able to wick away the heat!!! The Dip in lat April was not due to thermal shutdowns, but system maintenance. I also stopped folding after 5/15 as I was experimenting with how much I could affect our electricity bill by me shutting off the F@H client. The bottom out was in no way due to the CPU cooler. |
