Random Read/Write Speed

The four corners of SSD performance are as follows: random read, random write, sequential read and sequential write speed. Random accesses are generally small in size, while sequential accesses tend to be larger and thus we have the four Iometer tests we use in all of our reviews.

Our first test writes 4KB in a completely random pattern over an 8GB space of the drive to simulate the sort of random access that you'd see on an OS drive (even this is more stressful than a normal desktop user would see). I perform three concurrent IOs and run the test for 3 minutes. The results reported are in average MB/s over the entire time. We use both standard pseudo randomly generated data for each write as well as fully random data to show you both the maximum and minimum performance offered by SandForce based drives in these tests. The average performance of SF drives will likely be somewhere in between the two values for each drive you see in the graphs. For an understanding of why this matters, read our original SandForce article.

Desktop Iometer - 4KB Random Read (4K Aligned)

Random read performance is basically identical to the 520, and middle of the pack compared to other SSDs here. If we limit the comparison to just mSATA drives, only Micron's C400 does a better job.

Desktop Iometer - 4KB Random Write (4K Aligned) - 8GB LBA Space

Random write performance with data that's easily compressed is a clear win for the 525. It's clear the Intel has done some work under the hood as the 525 even outpaces the 520 here. Even when faced with incompressible data, the 525 does quite well (although its advantage over the 520 disappears).

Many of you have asked for random write performance at higher queue depths. What I have below is our 4KB random write test performed at a queue depth of 32 instead of 3. While the vast majority of desktop usage models experience queue depths of 0 - 5, higher depths are possible in heavy I/O (and multi-user) workloads:

Desktop Iometer - 4KB Random Write (8GB LBA Space QD=32)

SandForce drives do a good job scaling with queue depth and we see the 525 behave as expectedly here. Both it and the 520 converge to the same performance levels.

Sequential Read/Write Speed

To measure sequential performance I ran a 1 minute long 128KB sequential test over the entire span of the drive at a queue depth of 1. The results reported are in average MB/s over the entire test length.

Desktop Iometer - 128KB Sequential Read (4K Aligned)

Desktop Iometer - 128KB Sequential Write (4K Aligned)

Sequential performance is once again near identical to the 520, with the 520 pulling ahead in sequential writes.

AS-SSD Incompressible Sequential Performance

The AS-SSD sequential benchmark uses incompressible data for all of its transfers. The result is a pretty big reduction in sequential write speed on SandForce based controllers.

Incompressible Sequential Read Performance - AS-SSD

Incompressible Sequential Write Performance - AS-SSD

Incompressible performance is very similar to the 520, although once again the 525 does pull ahead slightly.

Introduction Performance vs. Transfer Size
POST A COMMENT

21 Comments

View All Comments

  • vcorem - Wednesday, January 30, 2013 - link

    http://m.tomshardware.com/news/Mushkin-480GB-mSATA... Reply
  • IanCutress - Wednesday, January 30, 2013 - link

    Mushkin uses a stacked daughter board to achieve 480GB. This is usually ok in a motherboard, but not in a z-height limited mobile device. While it's still electrically mSATA, it is technically outside the mSATA specifications which limit z-height (if I recall correctly). Reply
  • lukarak - Wednesday, January 30, 2013 - link

    What is the chip density in the 768 GB SSD board in a rMBP? Does it have 12? Reply
  • Kristian Vättö - Wednesday, January 30, 2013 - link

    Yup, the 768GB SSD in rMBP has twelve 64GB (8x8GB) packages.

    However, SandForce has much stricter restrictions, the SF-2281/2 can only access up to 64 NAND dies (that's up to 512GB with 8GB dies). Samsung could build a 1TB drive if they wanted to, they just don't see the market for it (yet).
    Reply
  • lukarak - Wednesday, January 30, 2013 - link

    Thanks for the answer.

    On a slightly related note, the iPad (late 2012) 16 GB, has a single NAND chip, as could be seen in the teardowns, and there doesn't seem to be any room for more. Do you perhaps know, or care to speculate, how the higher capacity ones are configured, especially the new 128 GB version?
    Reply
  • Kristian Vättö - Wednesday, January 30, 2013 - link

    The 128GB model is most likely using new 128Gb (16GB) NAND dies to enable the higher capacity. You can only stack up to eight dies in a single package, so a higher capacity die was needed before you could go over 64GB (8x 16GB is 128GB).

    At least Samsung and IMFT have 128Gb dies in production but they most likely weren't available in volume when the iPad 4 was initially launched, hence the delay.
    Reply
  • lukarak - Wednesday, January 30, 2013 - link

    Makes sense, thanks for the insight. Reply
  • SAMSAMHA - Wednesday, January 30, 2013 - link

    hi, Anand

    I am curious what board are you using to test this since none of the desktop board with mSATA that I know has SATA 6Gbps interface.
    Reply
  • SodaAnt - Wednesday, January 30, 2013 - link

    There are msata to sata conversion boards that work fine because msata and sata are electrically the same. Reply
  • Meaker10 - Wednesday, January 30, 2013 - link

    The msi gt60 can ship with a pair of msata gen3 slots configurable in raid. Reply

Log in

Don't have an account? Sign up now