User Rating: Be the first one !

9 comments

  1. blank

    > If you look close, you will see the performance listed at 550MB/s read with 520MB/s write

    Those specs are only for the 240GB version. See Footnote 1:

    “Performance values vary by capacity.”

    MRFS

  2. blank

    I agree…no mention of the SF brand…pretty lame. Imo the delay until Q1 2012 is most probably cuz there are still some issues to be worked out beyond just the bsod issue. We all know that SF’s has always been full out-of-box compatibility…even the SF1200’s have left over issue to this day. I suspect that this extra time before release will be spent by LSI’s engineers blitzing the fw to iron out the last bits and then they can uncork some of their hellified qualification ability. No way Intel is coming to market with the current state of the drives/fw…ofc all this just an edjumicated guess. 😉

  3. blank

    Excuse the typo….should be..

    We all know that SF has always fallen short of full out-of-box universal compatibility…

  4. blank

    Imagine how the entire IT industry would HOWL BLOODY MURDER
    if DDR3 SDRAM “write” performance were proportional to capacity.

    Let’s see, to drive this point home with an extreme example:

    10 CONTINUE

    32 GB @ 4 GBps
    16 GB @ 2 GBps
    08 GB @ 1 GBps
    04 GB @ 500 MBps
    02 GB @ 250 MBps
    01 GB @ 125 MBps
    512 MB (your abacus may be faster)

    We have “firmware” that will unlock these “throttles”,
    but you’ll have to pay more for the latest firmware.

    Now, repeat cycle with rotating platter storage: GO TO 10

    Repeat cycle again with USB 3.0 thumb drives: GO TO 10

    (Footnote 1: if your USB thumb drive is smaller than 16GB
    you can replace the “3.0” with “1.0”.)

    Can you say “oligopoly” with a straight face?

    HAVE I MADE MY POINT?

    20 GO TO 20 /* the longest program in the world, right here!

    MRFS

  5. blank

    With SSD controllers it’s all about parallelism. Most have 8 channels/16 bit lanes…for best results all 8 need to be utilized at once. Todays affordable nand memory chips are now up to 32GB or higher each. So a 128gb ssd would only have 4 nand chips leaving half of the controller totally unused…that’s why you see low performance with smaller drives. In the beginning nand chips were 8 or 16gb thus using all channels BUT the smaller nand is waaay to expensive cuz it’s not cost effective now days. Heh…you think it sux with SF, wait till you see the performance gaps for Ocz’s upcoming Octane…you have to get a 512gb drive to get max performance!! Nand prices are steadily coming down with each new process witch at some point will start to even out the size vs perf. ratio…but it may be another year or more…imo

  6. blank

    I do understand these properties of Nand flash. Thanks for the crystal clear summary: your understanding is obviously superior to mine.

    Nevertheless, I was reacting to the above recommendation that we “look close” and “see the performance listed at … 520MB/s write”.

    Well, I did “look close” and confirmed what you just explained so clearly below … only it was buried in microscopic text in “Footnote 1” ;(

    MRFS

  7. blank

    Keep in mind that there is a 120GB Intel SSD 320 Series drive, so it’s not necessarily a SandForce controller. SF based drives have 128GB of NAND in them, so it’s basically just a matter of how you want it to be reported. Other drives report the total capacity, although some of it is used by over-provisioning.

  8. blank

    LSI is a very promotional management team, they have for years promised wins in the hdd controller space that never come to fruition. when they announced the SF acquisition they made no mention or even a hint at a “new” ssd customer. in addition, while SF controllers tend to be the best performance on paper when new, they deteriorate quickly…i doubt intel would offer a 5 year warranty on a SF drive. and finally, the existing 510 doesn’t use SF, and it too has “off” capacities (120 and 250). i don’t think the 520 being a SF drive is a done deal, in fact based on the above, i’d say the odds are at best 50/50, but i suppose only time will tell

  9. blank

    If this is true, then its fantastic news. I would much rather trust SandForce based Intel SSDs in a server than from other brands.

Leave a Reply

Your email address will not be published. Required fields are marked *