We have decided to perform comparison test of memory modules that are certified at 533MHz or more precise at DDR2-1066 standard. We provided for this test several Dual Channel kits with 2x1GB capacity. Choice for this module speed came as consequence of the fact that most motherboards support 1066 standard. Capacity of 2GB RAM memory became standard also so our decision was actually led by these facts.

 

We wanted to compare and determine what memory gives the best performances at default settings and of course, to determine overclock margin for each module. As you can see in table with specifications, tested modules have three the most popular chips lately.

For test rig we used Intel based platform with ASUS P5Q Deluxe motherboard. This motherboard is based on P45 chipset that works with no problems with memories certified at 533MHz without overclock of FSB. Whole test was done at default settings for FSB and CPU clock, so that results can be completely comparable. We didn’t use performances for overclocked memories because we wanted to avoid influence of higher clock for FSB bus on results (as memories couldn’t attain identical maximum clocks, so the FSB clock was different in each case). If we conducted tests differently, we could get incomparable results that aren’t influenced only by memory speed.

 

 

In standard GeIL package are two modules with orange coolers. This type of coolers GeIL uses on DDR2 modules from Ultra series for a long time. Coolers are glued and nicely fitted. As on box, as on memory itself, it is clearly underlined that these modules are certified at DDR2-1066 with 5-5-5-15 latencies. It is also specified that these modules need voltage form 2.2V up to 2.4V which automatically provides warranty up to 2.4V for these modules. During tests, we didn’t have any problems with memory when settings that manufacturer appointed were used. Problems occurred with voltages set higher than 2.24V for RAM memory – memory just didn’t work stable. Problem may be result of incompatibility of our motherboard with modules or the exact specimen that we tested was troubled. Overclocking of memory was very decent, we even managed to pull out entire 570MHz and we believe that we could do much better if there wasn’t that little problem that we mentioned earlier. We really believe in that because Micron chips need a bit higher voltage than competitors do. While we talk about chips, we must say also that according to all information that we got, GeIL uses on these modules, from Ultra PLUS series, Micron’s D9GCT chips that are by performances slightly below D9GMH and D9GKX chips. In EVEREST test is clearly that GeIl Ultra 1066+ has shortest access time which tells us that secondary latencies are more tightened than on all other tested modules.

 

 

 

Kingmax doesn’t have special series of DDR2 modules so we took the fastest modules from desktop DDR2 series that are certified at 533Mz with 5-5-5-18 latencies. These modules come without coolers and withal are the only memory without them on this comparison test. Manufacturer on its official site says that voltage should be set between 1.8-2.1V for stable memory on DDR2-1066. But when we looked at detailed instructions in PDF available on Kingmax’s site, it’s clearly specified that for DDR2-1066 mode is necessary to set voltage at minimum 2V. It’s quite obvious that secondary timings are little bit “loose” comparing to competitors, so memory showed in EVEREST test to be slightly slower compared to other modules. It’s also important to mention that this was the only memory on test that could achieve magical 600MHz, or namely DDR2-1200 memory frequencies. For this attempt we set voltage at 2.1V, which is by our opinion great result. After some thinking about this result, we realized that this actually isn’t that odd, regarding the fact that Powerchip chips don’t need too high voltages for achieving maximum overclock. We found that these modules deserve Approved InsideHW award thanks to achieved performances.