ADATA XPG SX900 256GB SATA 3 SSD Review – Expanded Capacity and SandForce Driven Speed


Crystal Disk Benchmark is used to measure read and write performance through sampling of raw (0/1 Fill/compressible) or random data which is, for the most part, incompressible. Many new SandForce Driven SSD owners who cant wait to test the performance of their SSD often grab this program and run a quick test, not realizing that they are testing with incompressible data rather than compressible data used in testing by manufacturers.  We have provided compressible (oFill) results on the left with incompressible (random data) results on the right.

blankThe XPG is looking stronger with every test.  Low 4k random writes are up there with 90MB/s but, even better, the high sequential write transfer speed performance above 300MB/s when testing in highly incompressible, or random data, was amazing.  This was totally unexpected as this was the point where we first thought we might see that trade off in performance for that extra 7% capacity gain.


Up until recently, AS SSD was the only benchmark created specifically for SSD testing and it uses incompressible data.  AS SSD, for the most part, gives us the worst case scenario in SSD transfer speeds while using SandForce Driven SSDs as they use compression in storage as discussed earlier.  Many enthusiasts like to benchmark with AS SSD for their needs.

blankblankOnce again, these are some of the higher scores we have seen from ‘SandForce Driven’ SSDs and, although the IOPS may appear to be a bit low, this is typical of AS SSD IOPS results.

blankAS SSD Copy Bench is a very realistic benchmark, in that, it simply measures the transfer speed and time it takes to transfer a typical ISO, program and game from one part of the disk to another.  These are excellent results. Lets take a look at the AS SSD Compression Test:



Over the last little while, we have been assisting with beta testing new benchmark software called Anvil Storage Utilities which is an absolutely amazing SSD benchmarking utility.  Not only does it have a preset SSD benchmark, but also, it has included such things as endurance testing and threaded I/O read, write and mixed tests, all of which are very simple to understand and utilize in our benchmark testing.

blankThe definite pluses seen in this benchmark include the quick disk access speeds as well as something that we don’t often see in SandForce Driven SSDs, this being strong read and write IOPS results as shown with the 68, 213 IOPS result at 4k0QD16 read and the 83,013 IOPS result at 4k-QD16 write. This could very well be the top drive we have reviewed in awhile.


  1. blank

    I don’t think it’s entirely unexpected that ditching RAISE would help increase speeds somewhat. After all, the drive isn’t having to write the necessary parity(?) data anymore to recover from a die failure. With that said, I’ve seen increased write performance from SF’s using the new reference 5.02a FW anyway.

    So it looks like you do get slightly better performance overall and more capacity in exchange for less failure tolerance.

  2. blank

    It would be interesting to run a pair of these in RAID 0 to see what the low RAISE does to performance.

  3. blank

    Can’t see how there would be any major surprises, Crucial have been using the 0% over-provisioning for years, The only over-provisioning they’ve had has been the difference between the size of actual memory modules ( gigabytes ) and the actual size of the ssd drive ( Gibibytes ) it still leaves 7% difference for over-provisioning, which is classed as 0%.

    More over-provisioning still makes for a better ssd drive where it matters, but I guess size is everything as far as the market is concerned.


  4. blank

    This is a great review you have there as usual 🙂

    I am Muhammad Al-Jawhary, an Arab reviewer and technical editor, and I have a 128GB sample of the drive that I have reviewed here:

    That makes us two sides of the same coin, you have done the only review in English (as far as I know), and I have done the first and only review in Arabic, too 😀

    The performance is great, and it has 8 GB more than its brother; the S511.

    • blank

      Ummm isnt that my photo of ‘all the ssds’ in your review? I would appreciate the link back to our site and article if you are using any of our literature or piks.

      • blank

        Thanks for your reply, it is a great pleasure to be here 🙂

        Actually, it is your pic, and the pic has a hyperlink embedded (If you press the image, you will be directed to the great article).

        We may have some future cooperation together, I may be able to translate some of your guides and articles and post them in the Arab world, and of course, the credit is all yours, not mine.

      • blank

        When I click on the pik, it brings me only to the pik…. Can we fix the link so it goes to the article? And yes, I am sure we can affiliate post for one another in the future.

      • blank

        Actually, the image has a hyperlink as I said before, but the site uses a script to show the images, which in turn hides the hyperlink. Nonetheless, I have added a credit line under the photo that contains the link. Sorry for the inconvenience 🙂

        Have you seen those numbers I got with the drive? They used asynchronous NAND flashes with the 128GB version! The same NANDs used with Kingston’s SSDNow V+200.

  5. blank

    There is a slight inaccuracy in your reviews which is (implicitly) already mentioned above, but which i’d like to clear up. Flash memory has raw capacities of binary powers, this one e. g. has a raw capacity of 256 * 2^30 = 274877906944 bytes = 256 “GiBi” Bytes. Vendors prefer to use decimal powers, as the numbers are slightly higher, so this SSD has a raw capacity of 275 * 10^9 Bytes = 275 GB. 7% of these are used for over-provisioning leaving the user with 256 GB = 238 GiB. Windows, again, uses binary powers to display drive capacities, so the latter number is what you see there. (No, NTFS formatting doesn’t swallow 18 GB of the drive capacity. 😉 Other Sandforce SSDs with 240 GB = 223 GiB user capacity use 14% of the raw flash capacity for over-provisioning.

Leave a Reply

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