AData Xtreme Edition DDR3 1600X 7-7-7-20

[Monstru]

http://forum.crazypc.ro/attachment.php?attachmentid=33450&stc=1&d=1222075075


Slowly, but certainly the moment for DDR3 to take it's rightful place in our computers is closing in. On one hand, the coming of Nehalem is getting closer and closer, and on the other hand, massive price cut-backs will be possible in the near future. All the aspects are favorable for the start of DDR3 domination starting with the last months of thsi year, so why not test some of the available kits on the market? Therefor, thanks to the kindness of AData , I had the oportunity to review one of their overclocking targeted kits, AData Extreme Edition DDR3 1600X (http://oc.adata.com.tw/1_product_detail.asp?pid=AD31600XU).

[Monstru]

http://oc.adata.com.tw/img_products/p_Xseries_pkg.jpg


The product

Unlike other kits,which use blisters or various plastic boxes, this kit comes in a very interesting box, and a pretty large one for a DDRAM kit. Usually I wouldn't pay attention to the way a RAM kit comes packed, but in this case I cannot stop observing the fine attention to details and the very good taste this mere box has to offer. The box opens in a way that reminds my of a jewel case, and inside, in two compartments dressed in black velvet, we can find the two memory sticks. For the artistic impression, AData surely gets a 10 grade in my book.

The black colored heatspreaders are standard size, assuring the compatibility with any mother board or unusual CPU heatsink available on the market. Once removed, we can see that we are dealing with a single sided kit, that uses Samsung IC's, which is an advantage related to the voltage needed for the maximum frequency, but also a disadvantage when it comes to working with low latencies. I have a small netagive observation at this points, the contact between the IC's and the thermo-adhesive band from the heatspreader is really poor, so more attention should be given to that area in the future.





Specs

AD31600X001GM(O)U
240Pin Unbuffered-DIMM Non ECC
2GB Dual Channel Kit (1GB x2
PC3-12800 (DDR3 1600)
7-7-7-20
1.75V~1.85V
Lifetime warranty




Pictures


PIC1 (http://forum.crazypc.ro/attachment.php?attachmentid=33458&d=1222081717)

PIC2 (http://forum.crazypc.ro/attachment.php?attachmentid=33459&d=1222081717)

PIC3 (http://forum.crazypc.ro/attachment.php?attachmentid=33460&d=1222081717)

PIC4 (http://forum.crazypc.ro/attachment.php?attachmentid=33461&d=1222081717)

PIC5 (http://forum.crazypc.ro/attachment.php?attachmentid=33462&d=1222081722)

[Monstru]

http://forum.crazypc.ro/attachment.php?attachmentid=33463&stc=1&d=1222081780


Testbed


Intel Core 2 Duo E8400
Asus P5E3-Premium X48
AData DDR3 1600X - 2.0v, 1:2, 2T
Cooler Master Hyper Z600
eVGA 9800GX2
WD Raptor 74GB
Cooler Master UCP1100W

WIN XP SP2
Everest Ultmate Edition 4.50.1330



Testing

Keeping in mind that these modules are overclocking oriented, the goal of our test is to check the overclocking capabilities of this memory kit. Doing this, I have looked for the highest clocks possible, using CAS 7, 8 and 9. I didn't forget about those who are interested in daily use, so I also tested the maximum stable clocks, reached using the timings mentioned earlier. For this, I used Orthos Blend. Also, using the stable clocks, I tested SuperPI and Everest Memory Benchmark to see how the setup is performing using this DDR3 kit. Please keep in mind that Everest results differ from one version of the software to another, so not all results are comparable.

[Monstru]

Results

As I was saying earlier, this kit uses Samsung IC's. The characteristics of this IC type results in the way in which this kit behaves, and that is easy to see from the beginning. The voltage specified by the manufacturer for 1600MHz 7-7-7-20 is situated between 1.75v and 1.85v. However, I could use this settings Orthos stable with just 1.70v, which is a very good behaviour. In my opinion, this kit might have more to offer then what we see in these tests, because the motherboard plays and important part in reaching the maximum clocks, and maximum clocks is what this kit is all about. From what I have tested, this kit does not scale at all with voltage beyond 2.0v vDIMM, so I used exactly this voltage for today's test.



7-7-7-21 - 1860MHz CpuZ (http://forum.crazypc.ro/attachment.php?attachmentid=33464&d=1222081862)
7-7-7-21 - 1800MHz Orthos (http://forum.crazypc.ro/attachment.php?attachmentid=33469&d=1222081881)

8-8-8-21 - 1900MHz CpuZ (http://forum.crazypc.ro/attachment.php?attachmentid=33465&d=1222081862)
8-8-8-21 - 1840MHz Orthos (http://forum.crazypc.ro/attachment.php?attachmentid=33472&d=1222081881)

9-9-9-24 - 2060MHz CpuZ (http://forum.crazypc.ro/attachment.php?attachmentid=33466&d=1222081862)
9-9-9-24 - 2000MHz Orthos (http://forum.crazypc.ro/attachment.php?attachmentid=33475&d=1222081897)



I also used the opportunity to test the way in which X48 responds to different timings, so I tested the platform in SuperPi and Everest, for the same CPU and RAM clocks, but using different timings.


http://forum.crazypc.ro/attachment.php?attachmentid=33476&stc=1&d=1222081897

[Monstru]

Final thoughts


Well, this kit really does not like tight timings, which is not a very good thing for benchmarks. The clocks I was able to get using <9 timings are not at all impressive. However, I think that this kit is a very good option for a high-end gaming platform, because of the low voltage needed for reaching 2000MHz stable. Also, these are really cold memories, they barely got a little bit worm during testing with 2v. The major advantage when compared to DDR2, besides the compatibility with future systems, is the high bandwidth and the configurations options for CPU speed / divider / ram speed. Whatever CPU clock or FSB we will use, the RAM will not limit the system's performance, and the bandwidth and performance will be constantly at a high level, even if the timings can't be tighten very much at high clocks.