Test Setup

Performance & Overclocking

Mushkin has programmed the DIMMs with three SPD timings. The first to adhere to JEDEC standards at 266 and 400 MHz, while the last one is an EPP profile. As you can see the memory is set to run at 4-4-4-12-2T and 2.05V.

The first thing that was done when benchmarking this kit, it was set at the intended CL rating and the advertised voltage of 2.0V. Running the memory at this voltage worked flawlessly. To keep a common base, the CPU multiplier was lowered from 7x to 6x and the CPU to memory divider was kept at 1:1. Using a 400 MHz FSB means that the memory runs at the intended 800 MHz, while the CPU is running at an healthy overclocked speed of 2.4 GHz. The CPU manages over 3.2 GHz, so there is still plenty of overclocking headroom.

Pushing the memory at this CL rating and 2.0V, it managed to go up to 864 MHz. A very healthy overclock at default settings. Raising the voltage, we did expect to get even more out of the memory. But even with 2.4V, we only managed to tickle a few MHz extra out of the kit. Looks like it does not scale well with voltage. The next step meant lowering the voltage to 1.8V. which is default value by JEDEC standards. The memory only managed 762 MHz, but raising it to 1.9V, we were able to break the 800 MHz barrier easily. It manages just basically the same maximum speed at 1.9V as it does at 2.4V.

The next benchmark run was done at CL3-3-3-10. Starting at 1.8V and 667 MHz at 1:1 divider, yielded a booting system, which was not quite stable. Lowering the speed down to 660 MHz then translated in a rock solid system. Between 1.8 and 2.4V are only 36 MHz. This means that the memory maxes out at 696 MHz with CL3. Looks like this memory does not hold any surprises at all, which is a bit dissapointing.

The last benchmark run was done at CL5-5-5-15. This is where it really shines. Starting with 1.8V, we managed to push it all the way to 1022 MHz. This is very respectable for a 4GB kit at default JEDEC voltage. Raising the voltage to 2.0V yielded very little improvements - too little to make any real impact. Oddly enough, raising it beyond 2.0V yielded an unstable system for anything above 992 MHz. So at high voltage, the memory does not nearly go as far as with 2.0V. Even setting the dividers higher, did not yield any better results. The mainboard is capable of delivering 500 MHz FSB, which would translate to a 3 GHz CPU speed, thus this is not the limiting factor as well.




As you can see in the graph above, the memory does not act ideally with increased voltage. This only reflects our kit and your mileage may vary.