Mushkin PC3200 Level II Dual Pack DDR RAM

Company: Mushkin
Product: PC3200 Level II 512MB Dual Pack
Street Price: $190 USD
Date Reviewed: September 1, 2003
Reviewed By: Joey Chao

Test Setup:

For our testing we use an Intel 2.4C processor. The reason for this is that it has a low multiplier. The lower multiplier helps to achieve much higher FSB's since we will also be overclocking the CPU. We have tested the CPU to run up to 260FSB, 100% stable. At 260 the memory would have to run at 520 MHz DDR, so this should be plenty of CPU to max out almost any memory. For our testing procedure, we used MemTest86 to check for stability and SiSoft Sandra 2003 for our benchmarks. Also if we say stable, it means that the memory was stable through endless loops of MemTest86. If the memory could not run through MemTest86 without giving errors, we didn't even bother to boot into windows and run the benchmarks.

Results & Overclocking:

First we started out at SPD values set in the BIOS and used CPU-Z to make sure they were being detected correctly. The timings according the CPU-Z were 2.5-2-2-6 instead of 2-2-2-6. Mushkin is already aware of timings not being read correctly and the information is posted on their website. A nice aspect of Mushkin's website is that they talk about boards, what can be expected of them and they do testing to make sure that their memory works. We were happy to see that the ASUS P4C800 is listed on their website. We set the values manually in the BIOS to 2-2-2-6 and then did our testing. At stock voltage (2.55V) the Level II memory was 100% stable through Memtest86.

Overclocking the memory was a breeze. For every 20MHz DDR we overclocked the voltage had to be upped by .1V. At 2-2-2-6 the modules made it to 460MHz DDR. This is a nice overclock considering how tight the timings are. Keep in mind that this is over PC3500 speeds and in a dual channel configuration. There are PC3500 modules available that are rated at weak timings but the Mushkin was able to almost make it to PC3700 speeds at tight timings, only falling short by 6MHz DDR.

Next we tried to loosen the timings a bit to see how much higher we could get. We set the timings to 2.5-3-3-7, which is still fairly aggressive but more mild than the SPD values. We went ahead and shot for 470MHz at 2.85V and the memory did just fine through Memtest86. This was the highest we were able to get at any timings. We were hoping to go much further but we are not that disappointed. Notice the small gain from 10MHz and the change in timings according to Sandra.

Overall the memory overclocked very well. We were impressed with the speed and timings that we were able to use. The modules really seemed to like the voltages that we were throwing at them. Our motherboard is volt modded and we have options up to 3.3V but we felt that the information would not be that important since there are not that many users who volt mod their boards. Perhaps with more voltage, we could have pushed closer to 500MHz since we were getting 20MHz per .1V. Maybe we will do some further testing to find out.