AnandTech Storage Bench 2011 - Light Workload

Our new light workload actually has more write operations than read operations. The split is as follows: 372,630 reads and 459,709 writes. The relatively close read/write ratio does better mimic a typical light workload (although even lighter workloads would be far more read centric).

The I/O breakdown is similar to the heavy workload at small IOs, however you'll notice that there are far fewer large IO transfers:

AnandTech Storage Bench 2011 - Light Workload IO Breakdown
IO Size % of Total
4KB 27%
16KB 8%
32KB 6%
64KB 5%

Light Workload 2011 - Average Data Rate

The 840 Pro shows a similar ~20% increase in performance over the 830 in our light suite. Corsair's Neutron GTX is more competitive, but nothing can really touch the 840 Pro here.

Light Workload 2011 - Average Read Speed

Light Workload 2011 - Average Write Speed

Light Workload 2011 - Disk Busy Time

Light Workload 2011 - Disk Busy Time (Reads)

Light Workload 2011 - Disk Busy Time (Writes)

AnandTech Storage Bench 2011 PCMark 7, Performance over Time & TRIM
POST A COMMENT

96 Comments

View All Comments

  • kensiko - Monday, September 24, 2012 - link

    Interesting!

    Can Anand comment on that? Or verify it?
    Reply
  • jb510 - Monday, September 24, 2012 - link

    Same question. How do the Samsung provided numbers comparing the 830 to 840 on page 1:
    Active Power Use 0.24W 0.068W
    Idle Power Use 0.14W 0.042W
    Jive with the actual test numbers?
    Reply
  • Guspaz - Monday, September 24, 2012 - link

    I must be missing something, but on the first page, the rated power usage is:

    Active Power Use : 0.068W
    Idle Power Use: 0.042W

    While the actual measured use is 0.31W for idle and over 3W for active. That's about 50x higher, what's going on here?
    Reply
  • btb - Monday, September 24, 2012 - link

    Yeah, and no measurement of read power usage. Even though the typical user probably do an order of a magnitude more reads than writes. But thats a general anandtech testing problem, too much focus on the write specs instead of the more important read specs. Reply
  • ssj3gohan - Monday, September 24, 2012 - link

    See my comment above here: Anandtech measures the power consumption wrongly. DIPM is not enabled, which causes ridiculously much higher power consumption. Reply
  • krumme - Monday, September 24, 2012 - link

    The Anand hd bench was an excellent tool for comparisons, and one of the great tools Anand provided for us for delivering bm that reflected real world usage.

    What i derived from the first 2010 results was that the samsung disk was bad, but not so bad as it was portrayed by the reviewers and Anand. They were not worth their money, but they were slammed way to hard. At that time if we look at fx. strartup times, they were practically as fast as the fastest, so the criticism was ok, but a tad overblown.

    At that time many producers was as Anand said, not focusing enough on random 4k, but i can say for sure, nothing but random 4k write was usable here at anandtech at that time too :)
    Reply
  • sheh - Monday, September 24, 2012 - link

    As usual, I'm curious about the data retention properties of drives, and how they change across a drive's lifespan as P/E cycles increase. It would be difficult to check directly, but perhaps the manufacturers can reveal the details? I also wonder if TLC will be disastrous in this regard. Reply
  • Kristian Vättö - Monday, September 24, 2012 - link

    I asked about that in the Q&A session we had with Samsung engineers today but unsurprisingly, they couldn't tell us specific numbers. I hope to get my hands on 120GB 840 and 128GB 840 Pro to see what's the write endurance like, though that will take a while to test. Reply
  • FunBunny2 - Monday, September 24, 2012 - link

    It looks like even the Pro has no standby cap? How can that be anything better than a toss away consumer drive? Reply
  • sheh - Monday, September 24, 2012 - link

    They couldn't/wouldn't give numbers on P/E cycles or retention time?

    If you have a secret AnandTech stash of P/E exhausted drives (or could create one :), how about checksumming everything and testing every few months?
    Reply

Log in

Don't have an account? Sign up now