Setting this heatsink up was a pain from the very beginning. Our first task was to dislodge the sink from the included bracket, and considering that I'm used to the simplicity of the stock sink's ratchet design, I wasn't expecting to have to enlist the aid of a pair of pliers to do this task. Since our test motherboard (DFI NF4-DAGF) has been retrofitted with a standard metal backplate, we didn't need to enlist the aid of Spire's. Installing the sink was a whole new horror story in itself. We eventually found that it was best to put down the locking side first, and then use pliers to push the other side down. You can imagine our surprise when the force required to push the other arm down actually tore the screwdriver hole right out of the bracket, just narrowly escaping leaving our motherboard with gaping wounds.
Not cool spire. On top of this, it was exceedingly hard to press the plastic locking arm down, and considering how much force was necessary to just get this sink down in the first place, we wonder if this was even needed.
Our testing methodology is as follows - we take four separate measurements to show different load levels for the processor: We have a no load 15 minute period as our idle, 30 minutes of Folding @ Home to show a low priority load, 10 Loops of SiSoft Sandra's Burn in wizard to demonstrate a high priority load, and a standard S&M processor test loop to demonstrate a critical load. At this time, we utilize a less popular method of the motherboard's inbuilt sensors. Yes, we know this is less accurate to some respect, but it does however prevent misrepresentation of data due to a jarred thermal probe or strangely mounted heatsink. On top of this, we run two sets of tests - one for Stock (200*9, 1.45V) performance, with the processor generating a paltry 49W of heat, as well as an Overclocked (280*9, 1.65V) test where our processor puts out just shy of 100W of heat. Without further ado, let us proceed onto the tests.
Test System
- 3000+ Winchester Chip
- DFI NF4-DAGF
- 1.0 GB Crucial Ballistix DDR500
- NVidia GeForce 6800GT
- Ultra 600W Power Supply
- Open Air Test Bench
Basically, we can see that while the Spire sink starts with a small lead during the stock tests, it grows to a considerable one by the end of them. However, when the processor's heat output is ramped up, it meets up again with the stock sink and, with fan pumping furiously, manages to produce higher temperatures. We're sure that the soldered fins and rough base are the reason for this loss to AMD's offering, and we're also sure that if spire changed the design a little, it would be a winner.
