Sign in with
Sign up | Sign in

SanDisk Makes SSD up to 100x Faster

By - Source: Tom's Hardware | B 15 comments

The future of storage is heading away from the spinning magnetic disks inside most of our computers and towards solid state drives.

Besides the advantage of being much more reliable and durable, solid state drives (SSD) also have the advantage of being quicker to deliver data. SanDisk detailed today a new technology that it hopes to cement SSD as the performance choice, coupling it with new metrics.

SanDisk explained its new file system technology for SSDs, which it calls the ExtremeFFS, at WinHEC 2008 in Los Angeles and boasts that it could boost random write speeds by up to 100 times over existing systems.

The “FFS” in the brand stands for Flash File System (not the other use of the acronym) and the Extreme, of course, relates to SanDisk’s brand of premium flash products.

SanDisk appears to have geared its technology towards the needs of Windows Vista. ExtremeFFS operates on a page-based algorithm, which means there is no fixed coupling between physical and logical location. SanDisk explains that when a sector of data is written, ExtremeFFS puts it on the SSD where it is most convenient and efficient. This should result in an improvement in random write performance by up to 100 times in best case scenarios.

The memory company explained, “ExtremeFFS incorporates a fully non-blocking architecture in which all of the NAND channels can behave independently, with some reading while others are writing and garbage collecting. Another key element of ExtremeFFS is usage-based content localization, which allows the advanced flash management system to "learn" user patterns and over time localize data to maximize the product’s performance and endurance.”

Rich Heye, senior vice president and general manager for SanDisk’s SSD Business Unit, said, “This feature might not show up in benchmarks, but we believe it is the right thing to do for end-users.”

SanDisk products with ExtremeFFS are expected to ship in 2009.

Display all 15 comments.
This thread is closed for comments
  • -1 Hide
    jaragon13 , November 6, 2008 8:16 AM
    For Fuck's Sake!
  • 1 Hide
    lamorpa , November 6, 2008 11:39 AM
    They're going to deliver a 100X speedup, but it won't show up in benchmarks. Sounds like snake oil, don't it?
  • 4 Hide
    GlacierFreeze , November 6, 2008 12:13 PM
    lamorpa - "They're going to deliver a 100X speedup, but it won't show up in benchmarks. Sounds like snake oil, don't it?"

    Microstuttering doesn't show up on standard benchmarks. BF2 stuttering with only 1GB of RAM doesn't show up on benchmarks.
  • -7 Hide
    one-shot , November 6, 2008 12:23 PM
    ROFL 100X At 1st Comment!
  • 2 Hide
    tipoo , November 6, 2008 12:26 PM
    It's EXTREEEEEME!11!!

    Seriously, it's Extreme FFS. Oh FFS, it's Extreme!

    Someone should be shot for choosing such a name.
  • 0 Hide
    ceteras , November 6, 2008 12:59 PM
    I'm sick of this kind of new.
    Tell me you've got 100 faster SSD when they write at 6000MB/s.

    SSD = Same Shit every Day
  • 1 Hide
    Turas , November 6, 2008 2:14 PM
    Doesn't the OS need to support this new file system?
  • 1 Hide
    Anonymous , November 6, 2008 2:58 PM
    I think the biggest news here is:
    "ExtremeFFS incorporates a fully non-blocking architecture in which all of the NAND channels can behave independently, with some reading while others are writing and garbage collecting."

    I would venture a guess that this is the TRUE source of the'100x' performance increase.
  • 1 Hide
    KITH , November 6, 2008 4:16 PM
    this is the same thing that intel's drives do.

    it should have substantial performance increases.
  • 0 Hide
    falchard , November 6, 2008 5:23 PM
    When can I get my Laser Transistor Processor that promises to be 1000 fold faster then current processors?
  • 0 Hide
    Anonymous , November 6, 2008 6:13 PM
    You can get the promise from a marketing department that it is 1000 fold faster now long as you don't mind that it won't show up in bench marking
  • 0 Hide
    Area51 , November 6, 2008 6:56 PM
    What happens to MTBF?
    I think they are trying to get some P.R. before Intel drives Hit the Market.
    By the way, How come you don't have the SLC and the MLC offering from Intel listed in the Flash SSD Charts yet?
  • 0 Hide
    joex444 , November 6, 2008 8:15 PM
    Here's another poorly chosen acronym (though what could you reasonably do?) from Solid state physics, so it sort of applies. Weak Transverse Field. It is quite odd having your professor put this up on the board with a straight face...
  • 0 Hide
    efranchi , November 7, 2008 6:22 AM
    If it is, the SATA 2 and even the SATA 3 interface will be insufficient, and there will be only the PCI-E interface, like already used by some RAID combo (Raid card + cables + SSD in 2,5" SATA) who can already deliver 1250MB/s read 990MB/s write (with just 5pcs best SLC SATA 2,5 and the best RAID card) or some PCI-E raid card with "fusioned" on the card 80GB SLC (who deliver +/- 700/600 r/w sequential).
    Maybe in 2-3 years the motherboards will not have only 2-3 PCI-E for graphics card, need 1 more PCI-E slot for SSD ;-)
  • 0 Hide
    Anonymous , January 14, 2009 2:09 PM
    I think they are saying that with extreme ffs, you will see a 100x improvement in random write IOPS (likely 4k block sizes) They are not saying 100x increase in throughput. If you look at all non-intel SSDs today, they all suck donkey when doing random writes ... they just fail. Intel's drives fly along as fast as 10,000 IOPS for random write in some IOMETER results I've seen. This is an attempt at Intel's competition to catch up IMHO. For this reason, SATA 2 will be sufficient for 2009 ... especially for the laptops these are going to be used in.