The modules have a silver embossed “Geil” logo which looks quite nice; it’s a pity if you don't have a window in your case, or if your pc is on the floor, as the looks will be a bit wasted

So, let’s talk about the headtspreaders for one second. There has been a lot of discussion and articles about the effectiveness of heatspreaders on memory modules. Most people assume they are needed as the modules are being run at above averages voltages and settings. The evidence, from at least the testing I’ve seen, suggests that heatspreaders do aid with cooling of the modules but only slightly. In one article they tested the same module, overclocked, with both the heatspreader on and off, the presence in the heatspreader in the end gave a 10 MHz lead over overclocking witout the heatspreader applied.

On a different note, some people believe the heatspreaders give more protection from static electricity damaging the memory because of the chips themselves not being exposed. Personally I think unless you plan to skate round on your carpet on them, static isn’t really an issue.

I think Geil are wise to have placed heatspreaders on their modules, this way the user has the chance to push them slightly further than stock settings due to the extra cooling and because they look astounding.

Installation

As with any RAM modules they were a sinch to install, "Out with the old and in with the new." Below are a couple of pictures I took in the dar to emphasise how nice the blue heatspreaders look in a darkened case.

Testing

For the purpose of showing the performance of the memory I will put them through their paces I shall test them with some of the more popular benchmarking utilities available. The tests will be conducted at stock settings then at overclocked to show the maximum overclock obtainable and the performance increase this gives.

The test setup was as follows:

Amd 64 3000 Winchester
Abit AV8
Leadtek 6800le @ 12/6

For the duration of the tests I ran these modules at 2.8v

First Test

Super Pi mod 1.4 (1M)

As you can see the performance increased as I upped the memory speed. Strangely I hit a wall at 215 MHz. I was expecting more from the modules, but of course super pi might just be sensitive to overclocking. We shall see from the next tests if that is true

Second Test

Everest Home Edition Memory read test

Here things went exactly as expected, but the overclock was stable much further than I expected. The read bandwidth increases all the way to 250MHz! without a downfall but then becomes suddenly unstable. I have to say that around the 230MHz mark the system was getting unstable but the benchmark continued regardless. Also of note is the huge gains from overclocking the memory at 245MHz over stock a total of 1243MB/s extra bandwidth is obtained.

Third Test

Everest Home Edition Memory write test

As with the last test, huge bandwidth gains with sudden instability at the 250MHz mark.

Real World Benchmark

Counter Strike Source (1024 x 768, High settings, 2x AA)

To see what the modules performed like in a real world situation I used CSS’ benchmarking utility.

As you can see again performance increased with higher speeds, until around the 225 mark where the average fps actually drops with the memory 5MHz higher. This will be due to the instability affecting performance. Then testing at 230MHz gave fps drops in the region of 30FPS every 10 seconds followed by the game crashing the desktop showing high instability at these speeds.