Foreword
At this time, unlike the situation in the Russian market with a retail video cards secondary and higher price ranges, the situation with the proposals and the cost of memory is certainly gratifying DFI. Judge for yourselves: for about ten months ago, we submit to you the review and testing of twelve pairs of memory modules fiscal DDR2 standard PC2-4200, in which the average cost of the two sets of memory modules totalling 2 GB at $ 180! Today, the price of RAM declined to the point that just the same amount, recruited two budgetary modules only now standard PC2-5300 or even PC2-6400, can be purchased for less than $ 50. Impressive difference, is not it?

Because the total amount of RAM to 2 GB to this day remains a necessary and sufficient for most applications, including all the modern game, we finally decided to review the budget and testing of memory modules volume 2 x 1 GB standards PC2-5300 (667 MHz) and PC2-6400 (800 MHz) and provide the results to your attention. In tests we managed to get 11 pairs of standard modules, PC2-5300 and 3 pairs relating to the standard PC2-6400, from manufacturers such as Samsung, PQI, Kingston, Corsair, Aeneon, Hynix, NCP and the Patriot, part of which is shown in the photo below :

Memory came to us for testing parts, so do their common snapshot failed. So, the test length more than a month completed, the results obtained, so they turn to the study.

1. Test configuration tools and test methodology
However, before turning to the review of memory, describing their characteristics and study overklokerskogo capacity, it is necessary to say a few words about the methodology and testing of memory. Testing fiscal RAM conducted only in the dual mode of its work on the next block configuration of the system:

Motherboard: ASUSTek P5K Deluxe / WiFi-AP (Intel P35), LGA 775, BIOS 0603;
Processor: Intel Core 2 Quad Q6600 2400 MHz, 1.2875 V, L2 2 x 4096 KB, the FSB: 4 x 266 MHz, (Kentsfield, B3);
CPU Cooling System: Thermalright SI-128 + 120 - mm fan Scythe Minebea at 1140 rpm;
Termointerfeys: Arctic Silver 5;
Video: Sysconn GeForce 7900 GS GDDR3 256 Mb / 256 Bit, @ 575/1710 MHz (with Arctic Cooling Accelero S1 in passive mode);
Disk subsystem: SATA-II 500 GB Samsung HD501LJ, 7200 rpm, 16 Mb, NCQ;
Drive: SATA-II DVD RAM & DVD ± R / RW & CD ± RW Samsung SH-S183L;
Housing: ATX ASUS ASCOT 6AR2-B Black & Silver (at vduv and blowing installed 120 - mm hull fans Sharkoon Luminous Blue LED 980 rpm on the sidewall 120 - mm fan GlacialTech SilentBlade GT12025-BDLA1 at ~ 940 rpm) ;
PSU: Enermax Galaxy EGA1000EWL 1000 Watt (135 - mm fan at ~ 850 rpm and 80 at vduv - mm fan ~ 1650 rpm at blowing).
For cooling module was installed special cooler for the memory CoolIT Freezone:

Memory cards installed in pairs only on the first and third slots on your motherboard.

Tests conducted in the operating system Windows XP Professional Edition SP2. To test the stability of the functioning of memory has been used to disperse the following programs:

test program OCCT (OverClock Checking Tool) version 1.1.1b at its highest burden on the memory of a 35 - minute testing period from which the first and last time was 4 minutes of idle time and temperature stabilization system;
Further results refined with the help of "Torture Test" of Prime95 version 25.5 mode hour test;
final "cut" the memory of the four tests was from S & M version 1.9.0b-line "rule."
No doubt that even such a tripartite Stability Control BFG memory does not give 100% - yu guarantee full stability when overclocking. Nevertheless, the use of three different programs brings the likelihood of mistakes, or worse, failure to disperse memory to a minimum.

As for the tests, we were given a very large number of memory, a fairly lengthy period of testing, at this time exploring the possibility of reducing the memory subtaymingov were launched. According to tell the truth, the ultimate effect subtaymingov memory subsystem performance is low and marked only Benchmark such as Super PI or WinRAR. Thus, the memory tested in two major combinations timings: 4-4-4-12 and 5-5-5-15. The "Command Rate" has been set to "2 T". S taymingami 3-3-3-9 modules have not been tested, as well as the frequency at which the budget memory can operate stably at such taymingah too low to modern standards and, if necessary, to disperse the high frequency processors literally useless in practice.

In the motherboard BIOS parameter "Transaction Booster" has been to "Disable", it podparametr "Relax" in the value of "0". Also, the value of "Disable" was recorded and the presence in the motherboard BIOS setup option "Static Read Control." All of the settings contribute to the work of a maximum frequency of RAM to disperse. All other settings are left without changes in the values of "Auto". After determining the maximum memory operation to disperse at a nominal voltage of 1.8 V, last increased in the range of 2.0 V to the first, and then to 2.2 V.

Add that all prices of RAM and references to it are shown as at 15.12.2007 town Perhaps it is time now to move directly to the review of the study of memory and its test results.