Sign in with
Sign up | Sign in

Performance: IOMeter 2006.07.27

Tom's Storage Charts 2009: A New Test Environment
By

I/O performance depends on a drive’s ability to quickly reposition its heads, as well as utilizing the Native Command Queuing feature. NCQ organizes and executes pending commands in a way that delivers maximum performance by minimizing head repositioning activity. The tool we use was made for Windows, but it can also be executed at the command line, which is beneficial for bulk testing of several drives.

Unlike other benchmarks, which run a pre-set sequence of commands to determine performance, IOMeter is “programmable.” This means that you have to tell the benchmark what to do and in what order. Some years ago, we started to use IOMeter profile files, which simulate certain hard drive workloads, and we’ll keep using these four profiles for database, file server, Web server and workstation workloads. The only two additions are streaming read and write tests, which we introduced a few months ago.

Profiles


Read
Random
Block Size
Workers
Database
67%
100%
8 KB - 100%
4
Fileserver
80%
100%

512 Bytes – 10%

1 KB – 5%

2 KB – 5%

4 KB – 60%

8 KB – 2%

16 KB – 4%

32 KB – 4%

64 KB – 10%
4
Web server
100%
100%

512 Bytes – 22%

1 KB – 15%

2 KB – 8%

4 KB – 23%

8 KB – 15%

16 KB – 2%

32 KB - 6%

64 KB – 7%

128 KB – 1%

512 KB – 1%
4
Workstation
80%
80%
8 KB - 100%
4
Streaming Reads
100%
0%

64 KB – 34%

128 KB – 33%

256 KB – 33%
4
Streaming Writes
0%
0%

64 KB – 34%

128 KB – 33%

256 KB – 33%
4


The table looks complex, but is actually easy to read. The first column, “Read”, lists the amount of read versus write access. For example, 67% means that we are specifying about 2/3 read activity and 1/3 write access. The second column defines the percentage of random versus sequential access. 100% means that there is only random activity, which is the case for almost all server applications. The other extreme can be seen in the streaming reads and writes tests, where we want data to be read or written sequentially. The “Block Size” column lists the distribution of block sizes, which is used for each of the benchmark runs. Finally, the number of workers equals the number of threads used. Since we have a Core i7 quad core with Hyper Threading disabled, four workers seem to be best.

Example Results

Display all 28 comments.
This thread is closed for comments
  • 0 Hide
    kyeana , May 6, 2009 7:05 AM
    :D 
  • 0 Hide
    crisisavatar , May 6, 2009 7:51 AM
    woot charts ! now we need gpu ones ( i dont mind waiting til Q2 is over )
  • 1 Hide
    curryj02 , May 6, 2009 8:13 AM
    Loving the reinstatement of the 'article index' drop down menu... But I think someone needs to smooth out the rough edges. Minor points, but ones I will make nonetheless.
    First, the dimensions are such that you have vertical AND horizontal scroll... kinda annoying.
    Second, the 'index button' width is slightly smaller than the actual drop down menu that appears. So if you click the down arrow and move your cursor directly down (which because of the width issue, is not over the drop down menu) it deselects the index and it disappears. ARGGHHH
  • 5 Hide
    SpadeM , May 6, 2009 8:47 AM
    joeman42These charts are a disaster. The same exact label is used to denote multiple drives. E.g., Western Digital Raptor or Seagate 7200.11 are each repeated over a half dozen times on each chart. Trying to find a specific model requires you to follow the product link over and over again on each chart. I gave up, still not sure if the one I was interested in is even listed.....


    He's right, and if I select WD and Samsung as filters, and then choose a benchmark, I get all the HDD listed and i have to choose my filters every time I select a benchmark. The old chart system before the site was "pimped" was way better then this.
  • 0 Hide
    neiroatopelcc , May 6, 2009 8:52 AM
    The charts are bad, but the last version was bad too. The one before that was fine though.
  • 2 Hide
    xsamitt , May 6, 2009 11:30 AM
    I said we'd get harddrive review this week and lo and behold pappa was right.
  • 1 Hide
    acasel , May 6, 2009 12:44 PM
    I like the drop down menu now... Its much faster :-)
  • 1 Hide
    neiroatopelcc , May 6, 2009 12:47 PM
    acaselI like the drop down menu now... Its much faster :-)

    ye but I'd gladly trade the menu for the old layout with avatars and less gray on gray.
  • 1 Hide
    sublifer , May 6, 2009 1:39 PM
    Yay! drop down menu is back!
  • 1 Hide
    xsamitt , May 6, 2009 2:20 PM
    Yes but we were told we'd have our avatars back?i don't see them ,do you?
  • 1 Hide
    sandmanwn , May 6, 2009 2:23 PM
    yeah avatars would be nice to help break up the monotonous comment section. its just one big blob of text.
  • 2 Hide
    sandmanwn , May 6, 2009 2:26 PM
    WHAT HAPPENED TO OCZ DRIVES!!! Did Intel slip some money under the table?
  • -1 Hide
    neiroatopelcc , May 6, 2009 2:28 PM
    xsamittYes but we were told we'd have our avatars back?i don't see them ,do you?


    Jane said there was a chance, but she didn't promise.
  • 0 Hide
    fausto , May 6, 2009 3:16 PM
    there has to be a better way to do this. all i care about is real world performance. these charts are useless.
  • 3 Hide
    neiroatopelcc , May 6, 2009 3:32 PM
    The charts are useless if you don't know what you need. Yes. But they wouldn't be useless to most of us if we could see which model was performing how well. I know what I need to care most about is average read speed on all my drives except the system one, where access time is relevant as well.
  • 3 Hide
    stilespj , May 6, 2009 8:52 PM
    Ditto on the useless chart theme!!!
  • 0 Hide
    Anonymous , May 6, 2009 8:54 PM
    No numbers for Intel power consumption!???? what a joke. C'mon Tom's surely you can do better.
  • 2 Hide
    Area51 , May 6, 2009 9:01 PM
    I don't get two things.
    1. If this is a test bed then shouldn't you be using the fastest CPU available to you? Also I believe the i920 has a 4.8GT/s, so it can be a limiting factor when you are testing other components.
    2. Why are you not including the Intel SSD's They have been around for a while and they are still missing from your SSD charts.
  • 1 Hide
    Anonymous , May 6, 2009 9:01 PM
    No Intel SSD numbers at all! that's got to be the biggest oversight in the history of the universe.
  • 4 Hide
    drumerman , May 7, 2009 3:29 AM
    SSD charts need intel's drives as well as OCZ's vertex drives... these drives aren't that new, they should be listed
Display more comments