Home / Opinions / Memory Roundup / GeIL Ultra-X PC3200 Dual Channel2-5-2-2

GeIL Ultra-X PC3200 Dual Channel2-5-2-2

Price listed above is per 2x 512MB Modules.

Price Per MB - 20.6p

GeIL has arguably made the biggest splash in the memory market with some extremely aggressive prices coupled with very strong overclocking potential. The Golden Dragon series of DDR modules also won a lot of admirers with its distinctive looks and innovative bare wafer chips.

Ultra-X comes supplied in a very stylish, blue tinted hard plastic case with a slide-opening front panel. The modules are cloaked in anti-static bags and sit in holes cut in a foam packing panel.

Ultra-X is GeIL’s assault on the low latency market coming kitted out with - Hand Picked 5ns GeIL Chips rated to run at 200MHz at 2-5-2-2 timings. GeIL also claims to have optimised its SPD for dual channel operation, which sounds reasonable but which I can't easily verify.

Ultra-X comes equipped with a mirrored chrome finish, silver heat spreader that’s actually manufactured from pure copper before being tin-plated. Being copper it’s very efficient at absorbing heat from the memory chips but, as is common, the chips and heat spreader rely on a thermal tape interface, which however efficient, will impact on their effectiveness. No retaining clips are used in addition to the tape.

When it came to testing, the first thing of note was that setting the BIOS to run using the memory’s SPD settings did not set them at their rated 2-5-2-2 and I was forced the set the timings manually to give them a fighting chance.

Once set up manually the modules performed well at a stock 200MHz. Results were consistently on a par with other modules though slightly behind the Crucial sticks when taken across the full range of tests.

The 12-hour burn-in test completed first time and with no problems.

Overclocking the Ultra-X however was quite an experience. At 2-5-2-2 I hit the same 215MHz limit that the Crucial modules exhibited, but unlike the Crucial I was able to reach a heady 245MHz at 2.5-5-2-2. Slackening off again to 2.5-7-3-3 netted me an almost crazy 265MHz before finally at 2.5-8-4-4 I made it to an eye-watering 290MHz! To be honest, I had to back off to 285MHz to get the benchmarks to run for a prolonged period of time, but the fact that I never needed to increase the voltage to 2.9v or to lower the CAS Latency from 2.5 to 3 suggests it was the CPU that had reached its plateau not the memory, though at 290MHz I can’t imagine there was a lot more left.

For the record, increasing the voltage to 2.9v had no effect on pushing beyond 290MHz, nor did setting it to CL3.

comments powered by Disqus