- Comparing Water Coolers: We Follow Your Lead
- Thermalright's New Graphics Card Coolers
- Scythe Ninja Plus takes on Three TEC Heavyweights
- Radical CPU Coolers from CoolIT
- A Beginner's Guide For WaterCooling Your PC
- Vigor's Monsoon II TEC CPU Cooler
- Thermalright's HR-03 Is A VGA Cooler Gorilla!
- VGA Heat-Pipe Cooler Roundup 2006
- Six New AMD Coolers: Cold Enough For You?
- Readers' Responses to Strip Out The Fans, Add 8 Gallons of Cooking Oil
The Test Platform And Methodology - Four Cores Under Load
Source: Tom's Hardware US – Keywords: cpu, cooler, charts, 2008
Syndication:
The Test Platform And Methodology - Four Cores Under Load
We tested all of the coolers in this roundup on an Intel quad-core Core 2 Extreme QX6850 CPU. When all four cores are under full load, this model has a power consumption of 113.87 watts. In order to achieve this load, we use a multi-core version of Prime95 (ver.25.5a). The CPU displayed this power consumption when combined with the best water cooler we tested, reaching 62°C. The worse the cooler performs, the more energy the CPU will consume, as it will be operating at a higher core temperature. We decided to conduct the cooler tests on a quad-core processor with a high thermal dissipation, since many users overclock their processors, and coolers need to be able to satisfy these more sophisticated needs.

The cooler test setup.

The test platform with a water cooling solution installed.
The coolers were tested in a Gigabyte GA-G33-DS3R board with integrated graphics core mounted upright, the way it would be installed inside a tower case. Since most coolers these days utilize heatpipe systems to conduct heat away from the CPU, the position and orientation of the cooler can have a tremendous effect on its cooling performance, and by extension, its efficiency. Orientation may also impact the noise level of air-based cooling solutions. The temperatures of all four processor cores are measured directly at the digital temperature sensors (DTS). We calculate an average temperature and then create a line diagram from the data.
The duration of the measurement is determined by the maximum temperature that the cooler reaches. A very bad cooler may reach its highest temperature in a matter of seconds, while at the other extreme, a very efficient water cooler may take up to half an hour.

In this example diagram, the Zalman cooler has reached its final temperature of nearly 70°C after about 55 seconds.
- Previous page The Heat Is On - CPU Coolers Compared
- Next page The Criteria - Cooling, Noise And...
- 1 / 2
- Next
-
If my Artic silver 5 + Scythe Katana 2 cooler(for only a massive price of 25 USD! I can keep my outdated Pentium D under 100 degrees fahrenheit,infact around HALF of these tests,even under water cooling? Wow,according to this,I should actually worry about getting a quad core because of the heat.And no,I don't have any fans in my case,it's open,only fans are from my GPU,CPU,and my PSU.
No extra 4 250mm performance fans.And an X38? That's just unbelievable.
A good article, i especially like the tests for installation and sound, as those are what i would look into most when purchasing a cooler, and unlike CPUs, there are usually no charts to go along with.
but when installed in a normal ATX case, would be detrimental to the coolers performance. Especially coolers designed similar to the Noctua and Scythe Ninja plus.
I can't be certain about others but Arctic Cooling's Freezer 7 Pro is supposed to be installed like this(one would assume others of similar design would be the same)
http://bigrockies.com/media/cooler.jpg
As these have published manufacturer recomended installations why would you reverse it.
I recently read a review of the latest Noctua cooler at Legitreviews
where Noctua actually contacted them on this very issue... as a result they retested and found some improvement in cooling.
http://www.legitreviews.com/article/741/1/
As said the Noctua is setup with the fan on top blowing down, which is the only config that Noctua dont support.
The Thermalright has the fan in the middle blowing up so it looks like there is no real consistancy in the test methods... but I would like to hear from the testers in case they found some reason to use each particular setup.
- 1 / 2
- Next
-