Sign in with
Sign up | Sign in

Inside the X25-M SSD: 10-Channel Flash

Intel's X25-M Solid State Drive Reviewed
By

Why Is the X25-M So Fast?

The question is a good one: how the heck did Intel manage to create a MLC flash SSD that is faster than a high-end SLC product? And why do the drives store 80 GB or 160 GB, while silicon-based chips typically have capacities of 32, 64 and 128 GB? The answer is multi-channel flash. Intel uses its own SATA/300 controller and addresses ten different MLC flash channels at once, using a 16 MB cache memory. It also employs native command queuing (NCQ) to be able to distribute read and write operations across the available channels efficiently. Looking at the printed circuit board you’ll realize that the bottom carries the controller and the cache memory together with five of the ten memory channels (two flash chips each). The top side holds the other ten chips.

The current generation of 50 nm NAND flash stores 32 Gbit (4 GB) per chip. If you now use 20 of them (two per channel times 10 channels) you’ll reach exactly 80 GB. If you distribute reads and write across 10 channels it is obvious that performance will scale beautifully. However, we’re curious about how Intel implements the 80 GB (or 160 GB later on) on the 1.8” form factor, as 20 flash chips plus controller plus DRAM doesn’t fit into the 1.8” envelope (although the double-sided printed circuit board does). Intel will have to work with a smaller number of higher density flash chips, but we can only speculate about the layout. The specification of “up to” 250 MB read and 70 MB/s write performance indicates that not all models might perform alike.

The on-board 16 MB DRAM chip by Samsung is a part that we’ve also seen on various conventional hard drives, which we found somewhat amusing. This is the first flash SSD to implement NCQ, as it typically doesn’t make too much sense on a drive that allows direct access anyway. However, the DRAM buffer is not there to increase performance or to service NCQ for performance reasons. Rather, it is necessary to support write amplification control, which essentially is Intel’s attempt to improve performance while increasing life expectancy.

Display all 47 comments.
This thread is closed for comments
  • 0 Hide
    timaahhh , September 8, 2008 3:33 PM
    Very nice Intel, I couldn't drop that much cash into a harddrive, otherwise I'm sold :p .
  • 0 Hide
    customisbetter , September 8, 2008 3:49 PM
    200 Mb read solid. sweeeeeeet. i want one.
  • 1 Hide
    DXRick , September 8, 2008 3:53 PM
    Yummy! They greatly improved the write performance for a MLC drive.
    I would like to see it in a desktop compared to a VelociRaptor.
  • 5 Hide
    modtech , September 8, 2008 4:01 PM
    I look forward to the distant day storage devices are silent, last for a lifetime, contain no moving parts and perform like champs. We're nowhere near that day but it's coming closer one step at a time. :) 
  • 0 Hide
    Anonymous , September 8, 2008 4:03 PM
    It'd kill the velociraptor. SSD's are that much faster than regular magnetic HD's.
  • 4 Hide
    Anonymous , September 8, 2008 4:09 PM
    Yeah I'm surprised this review didn't include Velociraptor.
  • 0 Hide
    helopilot , September 8, 2008 4:38 PM
    Great review - Thanks! Request for future SSD reviews: please include the warranty period. SSDs are new technology and the length of the warranty is very important factor in my buying decisions.

    Agree with your conclusions: Intel has a killer product here. I *need* two of these - to go! :) 
  • 1 Hide
    Anonymous , September 8, 2008 4:38 PM
    Check out hothardware's review of these SSD's they did include the velociraptor.
  • -1 Hide
    Lavacon , September 8, 2008 4:46 PM
    modtechI look forward to the distant day storage devices are silent, last for a lifetime, contain no moving parts and perform like champs. We're nowhere near that day but it's coming closer one step at a time.


    Nontech?
  • 1 Hide
    DXRick , September 8, 2008 5:19 PM
    Quote:
    Yeah I'm surprised this review didn't include Velociraptor.


    They tested it in a laptop and thus only compared it to laptop HDs. However, given the latest articles about SSD for gamers and 14 SSDs compared (neither of which compared it to a VelociRaptor), I would think they would want to address those interested in using a SSD in a desktop.
  • 0 Hide
    NightLight , September 8, 2008 5:38 PM
    man this look sweet! i'm getting one! go intel!
  • 0 Hide
    4655434b20594f55 , September 8, 2008 6:48 PM
    Quantum Leap In Performance? in short yes.
    It is a small leap in performance. Maybe not as small a leap as the word quantum should describe.
    (Quantum Physics - a science of incredibly small things)
  • -2 Hide
    Area51 , September 8, 2008 7:00 PM
    Isn't Intel releasing SLC drives Also.. I thought this was the low-end MLC SSD Drive. If the Samsung part is a SLC then shouldn't we compare it to Intel's SLC also?
  • -1 Hide
    anon_reader , September 8, 2008 7:44 PM
    Ok...you guys just completed a "roundup" test of the fastest notebook drives on August 28. Why did you compare the X25-M to the SLOWEST of all the 7,200 RPM drives (the Seagate) you tested?

    My bet is that the WD Scorpio Black would have equalled or outperformed the X25-M in several of the applications benchmarks -- which would be the same result that IDC got in their benchmarks.

    And what's up with this "simulated startup" workload? Why on earth not test the actual startup (which, unlike your simulation, accurately tests synchronous IO capabilities). Again, in ACTUAL rather than simulated workload tests, these SSD's generally underperform the manufacturer's overblown claims. IDC's benchmark tests showed 7,200RPM HDD startup times faster than SSD. So...why "simulate" a startup workload?

    Finally -- why do the actual application benchmarks continue to show only marginal (and often -- MINISCULE) performance advantages for SSD?

    Based on the results of your 8/28 tests, if the X25-M had been compared to the WD Scorpio Black, the SSD probably would not have even come out on top in the applications tests.

    Looking at the application benchmarks, these flash-in-the-pan SSDs clearly have a long way to go before they can even reach across-the-board speedup of 2x over a fast HDD, much less meet the SSD hypesters ridiculous performance claims.

  • 0 Hide
    Master Exon , September 8, 2008 7:45 PM
    Yeah so how much?
  • 1 Hide
    anon_reader , September 8, 2008 7:55 PM
    Master ExonYeah so how much?


    Well, the X25-M scored 119 on SYSmark 2007 (overall) and the 'slowpoke' Momentus HDD scored 111. I think I want more than a measley 7% improvement before I'd (a) spend $700 and (b) give up 200GBytes of capacity.

    Don't you think?

  • -1 Hide
    anon_reader , September 8, 2008 8:00 PM
    anon_readerWell, the X25-M scored 119 on SYSmark 2007 (overall) and the 'slowpoke' Momentus HDD scored 111. I think I want more than a measley 7% improvement before I'd (a) spend $700 and (b) give up 200GBytes of capacity.Don't you think?


    Fyi...in case you missed it...it's at the bottom of the page:
    http://www.tomshardware.com/reviews/Intel-x25-m-SSD,2012-11.html
  • -2 Hide
    Anonymous , September 8, 2008 8:23 PM
    Sorry, not impressed. Performance only slightly better then conventional HDD? WTF are these guys doing.

    It seems like the SSD industry is looking to ONLY match the performance of the HDD where the technology should really shine well above it. Your telling me that whipping a read/write head over a disk spinning at 7200 RPM's performs comparable to reading and writing electrons directly out of a transistor? WTF? I am supposed to be impressed by this?

    If you can't read data off an SSD drive AT LEAST 4x faster then an HDD, don't bother me with it. The technology is not ready for prime time and the minor savings in power do not justify the tremendous cost per GB premium.

    Intel should be ashamed of even admitting making this drive. The whole SSD industry is a wash IMHO, this technology has been over promised and under delivered for such a long time I don't think the SSD industry knows what they are doing anymore. SSD should be cheaper, faster, and offer far greater storage capacities and near ubiquitous by this point in time after the promises made in the 90's.
  • 0 Hide
    Anonymous , September 8, 2008 10:04 PM
    Am I the only one that thinks what Intel is doing is merely providing a template for other companies to copy and sell them that template at a modest profit?

    And there is no contest that SSDs are a wash, right now. But this is an emerging technology that is going to be continually refined.

    Add to the mix Fusion IO's entry into the storage market (a flash pci express card) and SAS plugs being included on standard motherboards (some of the new x58s) we see a battle for the future of storage/hard drives and the removal of the bottle neck that has plagued computers for far too long.

    In my mind it's about time there was a serious push to remove the bottle neck of storage. Only the bleeding edge people are gonna be out a buck but how is that different from any other emerging technology?

    Go Intel for refining MLC tech and adding a controller.
  • 0 Hide
    asdasd123123 , September 8, 2008 10:59 PM
    1000 years mtbf? Did they use a time machine or what?
Display more comments