Wheter to find access times of WD hard drives?

Cerb

Elite Member
Aug 26, 2000
17,484
33
86
Not sure when (sic) they don't think it is an important parameter!
Because it's the same as everyone else's, it's kind of bogus, especially with NCQ being well-implemented, these days, and it tends not to change much, over time.
 

hasu

Senior member
Apr 5, 2001
993
10
81
Because it's the same as everyone else's, it's kind of bogus, especially with NCQ being well-implemented, these days, and it tends not to change much, over time.

Thanks for the reply

Does that mean that all WD 7200 RPM drives will have similar access times?

In fact I am wondering if and why I should go for WD2000FYYZ instead of WD2000F9YZ for a moderately loaded server?
 
Last edited:

Cerb

Elite Member
Aug 26, 2000
17,484
33
86
Thanks for the reply

Does that mean that all WD 7200 RPM drives will have similar access times?
All 3.5" 7200 RPM drives have about the same access times. All 2.5" 7200 RPM drives have about the same access times. It's also difficult to measure in a sense that matters, because now that they all have pretty good command queuing, average access times vary not only by drive, but by workload. Stuck at QD=1, most are around 15-20ms. But with higher queue depths, they can go into whole seconds.

In fact I am wondering if and why I should go for WD2000FYYZ instead of WD2000F9YZ for a moderately loaded server?
I like the cyan label better. :p The Re appears to have some vibration compensation, and an internal shock sensor. How much that matters, compared to what regular drives can do: your guess is as good as mine.
 

BFG10K

Lifer
Aug 14, 2000
22,709
3,000
126
To my surprise access times (seek time) are not mentioned in any of their spec sheets. Not sure when why they don't think it is an important parameter! Where can I find that info?
The reason is that there's no standard way to measure seek times, much like pixel response on displays.

Seek times can vary quite a lot across drives with the same RPM, even with the same platter configuration. As an example, Hitachi single platter 1TB drives are over 18ms while WD Blacks with the same configuration are only around 12ms. Even firmware revisions within the same drive can change seek times a lot.

IMO, the best measure of seek times is HD Tune's full disk test, and it's best to Google your drive (including firmware version where possible) to see if such tests have been done online.
 

hasu

Senior member
Apr 5, 2001
993
10
81
Hitachi single platter 1TB drives are over 18ms while WD Blacks with the same configuration are only around 12ms.
.....

it's best to Google your drive (including firmware version where possible) to see if such tests have been done online.

I could get very little information from third party testing/reviews, probably because they are newer drives. I wanted to know what WD would say about their own products.

These are the thrid party test results I could find:

14.3 ms for WD2000F9YZ
http://www.sotechdesign.com.au/western-digital-se-2tb-wd2000f9yz-hard-drive-review/

12.2ms for WD2000FYYZ
http://forums.vr-zone.com/hardware-...y-your-technology-related-purchases-1022.html
 

BFG10K

Lifer
Aug 14, 2000
22,709
3,000
126
That's perfect, 12.2 ms is almost the best you can get for a 3.5" 7200RPM drive (I've seen 11.8 ms from the older 2TB Blacks). 14.3 ms is starting to enter "average" territory.

Sequential performance appears to be about the same, so now you just need to weigh any other factors that are important to you.