CPU Water Block Round Up APOGEE GT/GTX vs. STORM vs. FuZion vs. APOGEE vanilla

Test bed Setup:

eVGA 680i nVidia reference board
Panaflo 86CFM fan used to blow over CPU area to cool MOSFETS and chipsets
Intel Core 2 Quad Q6600 @ 9x333MHz (2975MHz) overvolted to 1.6V (BIOS reading) to cause higher heat output
2x 256MB generic memory for basic system functionality
XFX Geforce 6200TC once again, this test bed doesn't need a high end graphics card
Maxtor Diamondmax9 120GB SATA
BenQ 1650 DVD+-RW

Water cooling components:

Swiftech STORM G4
Swiftech APOGEE
Swiftech APOGEE GT (bowed)
Swiftech APOGEE GTX (bowed)
D-TEK FuZion

(Arctic Silver Ceramique was used in order to avoid curing times associated with AS5)
Clearflex ½” tubing
Laing D5 at full speed
Thermochill PA120.2 radiator
2x SilenX 120mm fans running at 12V in pull configuration using the stock neoprene spacer
Home-made reservoir from a water bottle to simplify bleeding air bubbles
Pure distilled water with a little bit of iodine as a biocide.

Testing Methodology:

Each CPU block will be mounted 3x and the median result will be given. I feel that in order for this review to reflect the experience I believe a user will have with the block that an average temperature should be the one recorded in my results.

Core temp was used to record all temperatures. The CPU was left for 15 minutes to reach an idle reading, then was stressed using small FFT with Prime95 v25.1 to put a full load on all cores. The CPU was allowed to run at full load for 15 minutes to heat the water, then temperatures were logged for 10 minutes and averaged then rounded to get the most accurate results possible. The results must be reported to only 1 degree because the on-die sensor on Intel CPUs is only accurate to +- 1C

Ambient temperatures were measured by an indoor digital thermometer at the intake of the radiator. All temperature results will be reported in observed temperatures, because the room in which testing was done was kept at constant 22C.

The APOGEE GT and GTX were initially tested with a flat base, then with a bowed base, but due to their dismal results in my first round of testing, these numbers were not included. I was not able to “unbow” the bases completely, so I have left unbowed results out of my round up. I observed as much as 2C difference on the quad core in load temperatures when I initially moved from flat to bowed, just to give readers some idea of what to expect.