DDTUNG
01-14-2007, 07:30 PM
About a month ago I purchased a Swiftech H20-Apex Ultra Watercooling Kit through my good friend OPPAINTER with a view to cooling an overclocked Kentsfield rig that I was putting together for World Community Grid, a distributed computing project(DC) that utilizes the four cores 100% 24/7/365. As some of you might know, I switched from benching to DC a few years ago, and currently run about 100 PCs all year round. If benching is like a 100 metre sprint, DC is a marathon. Since the machines run unattended for the most part, absolute stability is required, or errors may be generated and affect the scientific results. Intel's Kentsfield cpu, when overclocked, is notorious for its heat output which puts most air coolers to shame, so I decided on watercooling for the two rigs that I plan to run. Then OPP asked me if I would do a review for Swiftech's new Apogee GT block which has been redesigned for the new generation of cpus with larger core area, and of course I agreed.
System specs:
Intel Q6700 Kentsfield ES
Asus P5B Deluxe(C1, bios 0910, no Vmods)
2 x 1G GSkill 6400HZ
Silverstone 560W PSU(38A 12V rail)
Tuniq Tower 120/Swiftech H2O-Apex Ultra WC Kit w/Apogee GT
Windows XP Pro SP1
The system was first set up using the Tuniq Tower 120, which is one of the best air coolers and the maximumn stable speed was 395 x 9 with 1.45 Vcore. 400 x 9 was stable most of the time, but would give an occasional error. Temps were as shown in Air.jpg, with 19.5C ambient. Please note that the test was done at 400 x 9.
Next I swapped in the watercooling gear and ran the system at 400 x 9 using higher and higher voltages, until I reached the maximum loaded Vcore of 1.58V on my stock P5B Deluxe. Temps were as shown in GT.jpg, GT1.5.jpg, GT1.55.jpg, and GT1.58.jpg, with ambient nearly constant at 19.5C. In short, the watercooling allowed me to run higher Vcore within the same temperature range. What good is that, you ask? Well, let's do some overclocking.
Into the bios and set the multiplier back to 10, the Vcore to give 1.55V under load, start with 360 x 10 and bump the FSB up by 5 at a time, and voila, stable at 375 x 10(GT1.55-3.75.jpg). It ran at these settings all weekend, proving stability. With 1.6 Vcore 380 x 10 would be easy. But that would have to wait for Part II, where I'll mod the P5B Deluxe and compare the Apogee GT to the Apogee.
DDTUNG:cool:
System specs:
Intel Q6700 Kentsfield ES
Asus P5B Deluxe(C1, bios 0910, no Vmods)
2 x 1G GSkill 6400HZ
Silverstone 560W PSU(38A 12V rail)
Tuniq Tower 120/Swiftech H2O-Apex Ultra WC Kit w/Apogee GT
Windows XP Pro SP1
The system was first set up using the Tuniq Tower 120, which is one of the best air coolers and the maximumn stable speed was 395 x 9 with 1.45 Vcore. 400 x 9 was stable most of the time, but would give an occasional error. Temps were as shown in Air.jpg, with 19.5C ambient. Please note that the test was done at 400 x 9.
Next I swapped in the watercooling gear and ran the system at 400 x 9 using higher and higher voltages, until I reached the maximum loaded Vcore of 1.58V on my stock P5B Deluxe. Temps were as shown in GT.jpg, GT1.5.jpg, GT1.55.jpg, and GT1.58.jpg, with ambient nearly constant at 19.5C. In short, the watercooling allowed me to run higher Vcore within the same temperature range. What good is that, you ask? Well, let's do some overclocking.
Into the bios and set the multiplier back to 10, the Vcore to give 1.55V under load, start with 360 x 10 and bump the FSB up by 5 at a time, and voila, stable at 375 x 10(GT1.55-3.75.jpg). It ran at these settings all weekend, proving stability. With 1.6 Vcore 380 x 10 would be easy. But that would have to wait for Part II, where I'll mod the P5B Deluxe and compare the Apogee GT to the Apogee.
DDTUNG:cool: