Noise & Temperatures
Testing was conducted using a combination of software and hardware with the Hardware Sensor Monitor program and our lab approved Fluke 62 Mini Infrared Thermometer. It offers the best accuracy in its class, +/- 1% of reading. It also has a very wide temperature range from -30º to 500ºC. HSM recorded temperatures over 60 minutes while FutureMark 3DMark 06 was ran on default visual settings with all tests enabled. This allows the GPU to ramp up and down like it would in a normal DirectX/OpenGL application where 3D action isn't always present on the screen or when users ALT+Tab out of their game environments to check IRC and come back. The Fluke 62 was used to backup temperatures at an interval of every five minutes to verify HSM readings at an ambient temperature of 25º C. Physically measuring the temperature without a thermal diode attached is an inexact science, and is something we hope to rectify in the near future. Overall the temperatures below suggest two things. One being the Cooler Master CoolViva Pro on average performs better than its competitor at it's own max and minimum RPM speeds at stock speeds. Our eVGA NVIDIA GeForce 7800GT at stock runs with a core speed of 445 MHz and a memory bus speed of 1070 MHz.


The second aspect one must keep in mind is the noise factor. At maximum RPM the Zalman VF900-Cu is uncomfortably noisy. This may not be an issue for some individuals who play games with headphones but while you aren't playing you are still susceptible to the high whine the fan is emitting. It's up to you if you want to dial the Fanmate every time you are in and out of game. If you have a motherboard capable of software temperature control then you're in luck. Otherwise, this writer would personally not want to sleep, watch a movie, or study for exams in the same room as a cooler this loud. Testing at the lowest and highest speeds provides no room for user error, therefore we also tested the slowest RPM on the Zalman cooler and found it to perform worse on average. The key thing here is that at its lowest RPM the Zalman VF900-Cu emits 18 dBA whereas the Cooler Master CoolViva Pro emits a constant 20dBA. This makes the CoolViva only a little louder, yet it still manages to perform significantly better. You might be saying to yourself "but Oscar, the difference between these temperatures is sometimes within two degrees, big deal!" and you are correct. These readings may even be within the inherent margin of error in measurement, but nevertheless managing to maintain a high temperature of no more than 62º C when the Zalman VF900-Cu had a high of 70º/77ºC is admirable.


Overclocking brought in basically the same results in terms of temperature difference. Since our testbed eVGA NVIDIA GeForce 7800GT already comes overclocked we found the maximum limits using both coolers to be identical. Overclocked core speeds reached 465 MHz while the memory speed topped out at 1100MHz. Attempts to go any higher would result in artifacts consisting of white random textures flashing across the screen while playing Valve Software's latest hit - Team Fortress 2. If you ever see this type of obvious errors that only occur after you've throttled up your GPU's clocks then it's a good idea to quickly stop and bring the speeds down before any permanent damage occurs. Don't rely on testing software to verify your overclock has passed. Your best bet is to get your frag on and see what a real world situation does to your card. Looking at the chart above shows the Zalman at minimum RPM struggling to keep the temperature down whereas the hardly louder Cooler Master CoolViva Pro generally does a good job. On average everything is about the same, and your own ambient temperature and computer setup more than anything will affect these numbers. Both products perform well for the most part with the Zalman VF900-Cu reaching slightly higher maximum temperatures. Which one should you go with? Well, the answer to that question depends on the rest of your system components.