Test Setup and methodologyLeeghoofd's I7 940 Test Setup |
|
|
|
CPU | Intel I7 950 @ 3.8Ghz 1.29 vcore HT ON |
Cooling | MCP655 vario 2 and 5, Thermochill PA120.3 |
Mainboard | Asus Rampage Extreme II 1204 bios |
Memory | 6Gb Corsair Dominator PC16000 C8-8-8-24 1T |
PSU | Enermax Galaxy 1kw |
Like in the
previous roundup we tested in the same way. Here's a quick overview
All blocks were mounted 3 times (in the by the manufacturers recommended optimal position ) by using the supplied mounting kit. This means no back plates or other stuff was used to improve temps (if the WB was delivered without back plate that is)
I used a little carton template to get the same amount of TIM on the IHS. The Arctic Cooling MX2 (provided by JAHA.BE) was spread out by the pressure of the water block mounting. After the tests, the TIM was checked for a good spread out. If there was a bad mount the test would be repeated.
Tubing used was 2 meters of Tygon Black 3/8" tubing.
The Pump test were done via the MCP655 Vario equipped with an EK Waterblocks D5-X TOP, this to allow 1/4 AG fittings (pump speed setting 2 and 5 tested)
The total amount of fluid used each time was 800ml of Aquatuning Innovatek Pro fluid.
All test were started at 22°C ambient temp. The rig was allowed to heat up the fluid during half an hour at idle speed 3.8ghz at 1.2Vcore (no EIST enabled). Small note : under load the Asus Rampage Extreme II, tends to overvolt. So during testing the Vcore was fluctuating between 1.21-1.22Vcore.
As testing program I used Prime95 Custom 20K run for 1 hour stressing all 8 cores ( HT enabled )
The CPU temp was monitored by Everest Home and the Core temperatures by Realtemp. The latter is a nice program as it logs the highest core temps measured (making my life easier).
Due to the new i950 CPU (instead of the i940) and the use of other thermal paste we opted to retest all the previous blocks.
Enough Chit chat (already glad if you read through it all) let's show some results.
”Low Flow” Pump Speed 2 Results :The Watercool blocks are sandwiching one of the NDA blocks I got from
Enzotech to try out. Sadly I can't release any further info or pictures on it, as I don't have had the green light from them yet. But you can see they got a nice block up their sleeve in the near future.
It's amazing to see how most of these different crafted/designed blocks are really so close to one another regarding performance. The Heatkiller is really THE block to own at the moment for the enthusiast crowd. There's no real alternative, if you are looking for that last half a degree. The new blocks really are all performing well in the low flow setup and are giving some of the previous high end ones a good run for their money. Even though it's not one of the better performing blocks in this roundup, the lightweight Phobya block is close to the EK Supreme and Swiftech GTZ block. The latter two being quite respected blocks within the water-cooling community.
I'm still puzzled about the performance of the Yellowstone block, being claimed to be specially designed for I7. Yet it not even manages to nibble of 1 degree off the older Livingstone design. Without taking into consideration the king of the hill Heatkiller and the NDA block, most blocks are within one degree and a half of each other. Talk about a close competition.
”High Fow” Pump Speed 5 Results :Once we whack up the pump speed, we see a big climber in the roundup. Namely the highly restrictive
Koolance 350 AC block gains some serious ground. The
EK Supreme also known for its restrictiveness, benefits pretty good too from the extra flow. This also counts for the
Heatkiller v3.0 block as it enlarges the gap to almost 2°C towards its nearest rival.
Like in the lower flow setup, most blocks remain within the same tight margin of each other. Plus for most the extra performance gain over the low flow setup is negligible. But there's absolutely no denying it, the
Heatkiller v3.0 is really a class on it's own. However
Swiftech just released their new
Apogee XT block (which is claimed to best their current GTZ block by 3°C under certain circumstances) A review sample has been requested and granted, so an update will follow shortly... if it can live up to the hype, it could be a worthy contender for the performance crown.
EK is also working hard on a new block or at least a new mount to improve performance... we will keep you updated as soon as any news pops up.
Comparing the two pump speed settings and the gain (if any) expressed in this chart:
Thank you!
Must have been hard work benching so many blocks... and doing them all again just due to change of thermal paste. Thorough to say the least...
Though I am new to watercooling this is the only article I have found that tests this many waterblocks. This will sure help me in buying the right block for me.
Are there any other sources available like this?
None the less. Just felt like after all that hard work you deserved at least one thank you .
And I hope that you will expand this article over time so that it could be a big buyers guide almost... Covering as many waterblocks with real tested data as possible? Or does this exist already? Would be a major way for madshrimps to become even more unique?