Memory Prelude

Intel designed Core i7’s performance around triple-channel mode and low latency DDR3-1066. In fact, with three channels of DDR3-1066 memory, the platform will provide up to 25.6GB/s of memory bandwidth and nearly 38.4GB/s at JDEC approved DDR3-1600 memory speeds. Intel figures the bandwidth numbers available at DDR3-1066 coupled with the on-die memory controller and relatively fast secondary cache speeds should provide enough performance to satisfy anyone. For the most part, they are right.

However, we are never satisfied so we set out to determine if applications we utilize everyday would benefit from increased bandwidth, improved latencies, or a combination of both. Obviously synthetic memory benchmarks improve greatly as you increase bandwidth and decrease latencies. The memory companies (with help from us at times) tout their latest and greatest memory products with these benchmarks while they are staples on the benchmarking circuit.



We were discussing our initial overclocking results with a couple of BIOS engineers and came away surprised. We learned they typically tune memory performance utilizing a combination of Sandra/Everest and Super Pi. That did not shock us but it explains why time after time our various test systems would perform exceptionally well in the synthetic benchmarks and then fall flat on their collective faces when faced with heavy multitasking situations. We believe too much emphasis is placed on trying to win a benchmark that in no way reflects the true performance of applications or enhances the stability of the system.

That brings us to another story. We had planned to incorporate a full overclocking section in this article but our DDR3-1866 and DDR3-2000 kits based on the Elpida DJ1108BASE, err Hyper ICs, have been experiencing technical difficulties as of late. As in, the damn things are failing faster than we can keep replacements in the labs. This is not a vendor specific problem, nor does it seem to be a voltage problem either, as we have had modules from various suppliers fail at stock VDimm or at 1.8V when trying to push through to the DDR3-2300 level. We have asked the memory suppliers a few pointed questions and are waiting for answers. In the meantime, replacements are arriving shortly. We will follow up with additional DDR3-2000 overclocking scenarios in the near future.

Index The Kits, Pricing and Designations
Comments Locked

47 Comments

View All Comments

  • BrianInfo - Thursday, July 30, 2009 - link



    1.) The recommended DDR3-1066 Patriot PSD36G1066KH is non-ECC and unbuffered. Will this be an issue for the personal desktop use, mainly for video/audio transcoding/muxing?

    2.) The article does not specify how to achieve the CAS5. Does anyone succeed the CAS5 with this Patriot DDR3-1066 and/or overclocked DDR3-1200?

    3.) Is it possible to achieve the CAS5 with other brand, such as GeIL CAS7 GV36GB1066C7TC or Crucial CT3KIT25664BA1067??

    4.) According to Tom's Hardware, not all the X58 motherboards support CAS5 http://www.tomshardware.com/reviews/cheap-x58-moth...">http://www.tomshardware.com/reviews/cheap-x58-moth...
    so does it mean I cant obtain the ideal performance with the combination of "ASRock X58 Extreme" motherboard with "Patriot PSD36G1066KH" Tweaked-CAS5 DDR3? I really like the Asus P6T SE but the layout of the two (blue) PCI Express 2.0 x16 slots, too close to each other is killing me:(


  • iwodo - Sunday, June 28, 2009 - link

    Spend the extra money on an SSD rather then faster Memory. Although i would love to see how Tri Channel differs from Dual Channel. ( Same results? )
    And how integrated Graphics differs in memory speed.
  • lemonadesoda - Friday, June 26, 2009 - link

    Gary, you obviously get paid by wordcount. Too much waffle. Get to the meat faster.

    And what about this $200 between cheap and expensive DDR3 kits? How about seeing WHERE that $200 could have been spent in OTHER upgrades, ie. CPU or GPU, and running benchmark comparisons against THAT setup vs. the first.

    Otherwise, thanks for running all the tests and creating the data for us to review :-)
  • PrinceGaz - Thursday, June 25, 2009 - link

    As soon as I looked at the chart on the first page of the article, it was clear that the DDR3-1600 C9 memory would represent the best bang for the buck. It was only marginally higher priced than the DDR3-1066 C7 (and the DDR3-1066 C5 doesn't count as you overclocked the DDR3-1066 C7 for that purpose), and was sure to perform at least as good, and probably better than significantly higher priced DDR3-1333 modules at C8, and almost as good as the DDR3-1333 C7.

    Given that faster memory modules only ever have a negligible effect on real-world performance (unless you are a pirate and spend a lot of time using PAR2 checkers to rebuild damaged files from newsgroups, or WinRAR to then extract the original files, the sort of stuff I certainly won't say I do), then you may as well get the cheapest brand-name memory available. If you're spending in the upper mid-range on your CPU, then go for slightly faster (like the 1600 C9) but still quite cheap memory-modules. The only people who should be getting those top-of-the-range modules or anything close to that price level are those who have already decided on the fastest Core i7 CPU avilable (i7 975 currently), as otherwise they're wasting their money on the wrong bit of hardware.
  • Souka - Thursday, June 25, 2009 - link

    I'm more confused than ever... what to buy?

    Putting together a new system...
    Win7 or Vista x64
    i7-920 CPU
    x58 MB
    ATI 4890 video

    Memory? I'd like to go with 6GB (3x2gb), thought I had my choices narrowed down to:
    Opt1: $138 mushkin 6GB(3x2GB) 1333(PC3 10666) Model 998706, Cas6, 6-7-6-18
    Opt2: $175 mushkin 6GB(3x2GB) 1600(PC3 12800) Model 998691. Cas6, 6-7-6-18

    But after reading this article, and your comments... I'm very uncertain....help!
  • PrinceGaz - Thursday, June 25, 2009 - link

    Given a choice between either DDR3-1333 or 1600 memory with identical timings (6-7-6-18), but a $37 price difference, I'd probably go for the 1333, especially as you have decided on an i7-920. If you want to spend a little more to improve yours sytem performance, the memory speed is the last place to look at to do so.
  • SiliconDoc - Sunday, July 5, 2009 - link

    I took just the opposite from it - he has a 4980 there - overclockable nicely, is saving quite a bit with lower cpu, and what else can get him another 2,5,7,14 % in framerates, especially minimum in some cases ? I think the higher ram is much better and worth it after seeing this articles results.
    He will OC the 4890 and the cpu likely, so that fast ram will give it that great finisging kick and make it awesome.
    Spend the $37 and be happy - definitely.
  • Hrel - Thursday, June 25, 2009 - link

    G.Skill 1.5-1.6v 1333 Cas 8-8-8-21 = 65
    Crucial 1.8v 1333 Cas 6-6-6-20 = 64 (sold out right now)

    OCZ 1.9v 1600 Cas 7-7-7-24 = 54 (after 20 mail in rebate)
    Patriot 1.9v 1600 Cas 7-7-7-20 = 95 (free shipping)
    Patriot 2.0v 1600 Cas 7-7-7-20 = 100 (Green, which I like)

    So basically, it really ONLY makes sense to buy the OCZ 1600 Cas 7 for 54 bucks, why pay more for less?!!!
  • sonci - Wednesday, June 24, 2009 - link

    Really boring for me,
    lucky, I didnt read it all
  • StraightPipe - Wednesday, June 24, 2009 - link

    cough * troll*

Log in

Don't have an account? Sign up now