Dance123

How fast is the new WD3200KS (with NCQ)?! Is access time slow?

Recommended Posts

Hi,

1/ Anybody here already has a WD3200KS drive?! How well does it perform?! Does the WD3200KS perform better or slower then the WD2500KS. According to a user review at Newegg the 3200KS has 107 gig platters where the 2500KS had 83 gig platters. Which one is best?!

2/ Second, anybody with a 3200KS can tell me the random access time. A Newegg user review said that it had only 14.5 ms, which is rather high, cause according to Storagereview.com review the 3200JD has 13 ms and the 2500KS has 13.6 ms. The 3200JD also has 100 gig platters, so I don't understand why the newer 3200KS would be slower. Anybody can measure the access time of the 3200KS to confirm if it is that slow. I expected it to be similar to the 3200JD = 13 ms and not 14.5ms which is slow. I do audio recording/production work, so access time is important. Anybody with a 3200KS can measure the random access time?!

3/ Is the 3200KS as silent as the 2500KS. Can the larger platter size have any impact on noise? Also, what type of noise does the 3200KS produce, like are there any annoying mid or high frequencies (whinning?) in the idle noise?!

4/ can you tell when will there be an offficial WD3200KS review so we can compare with 2500KS. The user reviews at Newegg are very positive but I would like to see an SR review to be sure.

Thanks very much for all good replies!!

Share this post


Link to post
Share on other sites

It doesn't really help as I don't work for SR nor do I even know how to properly test it, but I'll be getting the 3200KS tomorrow. I ordered it at ZipZoomFly for $139, and I hope it's worth it

Share this post


Link to post
Share on other sites
Guest 888

Generally higher-density platters are better. The most impressive effect is in sequential read/write speed which is always better with HD platters. So in that characteristic WD3200KS must be sure faster than WD2500KS.

There are some theoretical reasons why seek time can be slower with HD platters but real experiences with the latest newer drives have not confirmed this overally. I have no idea why WD3200KS must be slower seeker than WD3200JD. Their platters and drive mechanics must be pretty same.

In drive's noise characteristics the platter capacity (density) rarely impacts. The platters count impacts, but WD2500 and WD3200 have 3 platters both. In fact WD3200JB and WD3200JD are very silent drives. No high whine at all, only very weak low-frequency hum. But this may vary depending on individual drive from different manufacturing run or so.

I don't have WD3200KS yet, so I can't say anything exactly about it. But I am also supporting your proposal that SR would review this drive in nearest future because of just this WD's model seems to get pretty high attention and interest from various users now...

Share this post


Link to post
Share on other sites

I have a WD3200JB. While it doesn't have high frequency whine, it does exhibit some mid-high frequency "hissss" (depending on orientation of the head arm assembly). Idle noise resembles my 7K250 (which also makes the "hisss" when actuator arm is in certain direction), but when it comes to the seek noise, they are quite different.

HDS722525VLAT80:

no-AAM: random access time 12.0ms, audible crackling noise, some low-frequency rumble.

AAM: random access time 16.5ms, audible crackling noise.

(Almost no benefit enabling AAM in this drive.)

WD3200JB:

no-AAM: 13.4ms, possibly a bit lower frequency clicking, no mentionable low-frequency rumble. (A bit like 7K250 with AAM, but faster.)

AAM: quiet enough, haven't tried. Or possibly I have, I've just lost my benchmark results. (Shame on me.)

From the drives I've owned, FDB Maxtors are probably amongst the quitest when idle. They mostly just "hummm". The seeks, on the other hand, and absolutely terrible (and slow, DM+9 200GB: 13.6ms) compared to 7K250. Enabling AAM on Maxtors does help a LOT.

But is it really relevant how much noise they make when idle? It's mostly vibration that makes the noise audible. And measuring vibration is kinda hard, especially since there is many different types of computer cases and they tend to have different frequencys that make them vibrate. Even changing the slot where the HDD is mounted makes a big difference. (Magic word: decoupling.)

Share this post


Link to post
Share on other sites

Nobody at all who has the WD3200KS. I would really like to know about that slow average read access time since that it's important to me cause I do things like audio recording.

PS: please mention which tool you use to test access time. Which one is the most accurate by the way. Is HD Tach reliable cause some say it isn't. I believe Storagereview.com uses IPEAK SPT's AnalyzeDisk suite to measure Average Read Access Time.

Thanks very much!!

Share this post


Link to post
Share on other sites

Try HD Tach which you can find over here: http://www.simplisoftware.com/Public/index...?request=HdTach and please post the results here. It would be best to run it a couple of times to see if you get the same results.

Perhaps other people here can recommend other (better?) programs, but HD Tach is easy to install and use.

Edited by Dance123

Share this post


Link to post
Share on other sites

Winbench99 or AIDA32 perhaps? HD Tach is way more popular than it should be, considering the substandard quality of STR graph, etc.

Winbench99 makes more accurate STR graphs (graph is done making a sequential read of all sectors in a selected partition). AIDA32 makes a sequential read of either all sector on physical device, or "quick linear" option performs read for short blocks (like HD Tach does, but using less blocks and more interpolating). Because AIDA32 benchmarks physical device, the HDD can be benchmarked without partitioning or when it's partitioned into one or several pieces. Winbench99 works best only if the HDD has one partiotion. Otherwise it will only draw a graph out of selected partition and calculate the random access time within the selected partition. Combining multiple Winbench99 graphs requires photoshoppin and they are scaled into size and different partition are not scaled equally. Combining the random access times of HDDx partition 1 and HDDx partition 2 into random access time of the HDD as a whole is impossible.

AIDA32 also supports "linear write" option. This benchmark produced STR write graph but is destructible for all data (including partitioning information). Basically the linear write will zero-fill (or one-fill?) the entire drive, quite securely deleting remnants of data. (Of course random value seq. write is more effective, and to make it impossible to recover data even in cleaning room would require several passes.)

AIDA32 can also monitor SMART values and is capable of listing about all the components inside the PC case (unfortunately AIDA32 is no longer updated so newer computers have components that are unknown to AIDA32). HDD benchmark is just a addon, only a small fraction of it's functionality.

Winbench99 is primary intended for benchmarking. AFAIK, it cant read SMART, etc. Nor can it perform destructive write tests or acquire random access time for drives that are partitioned in two or more pieces.

I recommend installing both utilities. They are both useful. AIDA32 probably more useful as it has SMART reading capability.

Share this post


Link to post
Share on other sites

As for easyness of use, Winbench99 might not be ideal. The way the program saves the STR graphs is far from ideal. An example: Filenames for saving is asked before starting a several hours long HDD benchmark. Filename for CSV-file is also asked. If you intend to save the picture in X:\DIR\new_HDD.BMP and CSV in X:\DIR\new_HDD.CSV but leave the extensions off (assuming the program will add these extensions), it will save the BMP first (without the extension) then save the CSV (without the extension) overwriting the picture that might have taken several hours to produce. User-friendly. Also, if there's a bad sector somewhere on the benchmarked partition, Winbench99 refuses to save the picture, for some weird reason! Imagine the frustration of scanning the drive several hours, finding that the picture was not saved. Then run the benchmark again from start and take a manual screenshot just before benchmark ends. And, yes, I have cursed for hours using this software. Unfortunately there's no option if one want to make accurate STR read graphs of drives that have bad sectors. AIDA32 simply halts on the first bad sector and retrys it infinitely.

AIDA32 on the other hand, is very easy to use when you know where to look at. Heres how you find it: launch the program (AIDA32.EXE), click on button that says "Plug-in", click "AIDA32 Disk Benchmark". Disk benchmark opens in a different window. Instructions are under the "About" tab (that tab should be automatically selected when plug-in opens). Benchmarking options are linear read, quick linear read, random read, buffered read, average access and (if you enable write tests) linear write and quick linear write. The drive that's the target of the benchmark is on the lower left (dropdown menu).

Share this post


Link to post
Share on other sites

The following set of tests were done with NCQ off and AAM enabled. Without AAM enabled, the drive makes the loudest crunching sound during seeking I've ever heard. If for some reason theres a big demand, I can do it with NCQ enabled and\or AAM disabled, but I just wanted to bench it the way I would use it. Side question...are there any Windows programs that'll let me manipulate AAM so I don't have to use a floppy disk to get into Dos using Hitachis software?

HDTach:

hdtachaamenabled9kw.th.png

AIDA32:

RandomReadAAMEnabled.th.pngQuickLinearReadAAMEnabled.th.pngBufferedReadAAMEnabled.th.pngRandomReadAAMEnabled.th.png

Edited by aggybong

Share this post


Link to post
Share on other sites

Average access time only 17ms?!! Isn't that real slow!! Seems both tests you did gave that number so it must be correct.

You said you enabled AAM. Which setting did you use? According to wdc.com you can go from "Seek mode 0" (the loudest) to "Seek mode 3" (the most silent). Which mode did you use?

You said with AAM disabled seeks are really loud? Is that seek mode 0? Which drives have more silent seeks when set to mode 0

By the way, for setting AAM, what did you use? Can't you use Data Lifeguard Tools or Diagnostic downloads from wdc.com to make that setting?!

Thanks for the info!

Share this post


Link to post
Share on other sites

I used Hitachis tool as Western Digital has no software which lets you manipulate the AAM setting. I don't know what setting it's on, as Hitachi had a bar which ranged from 128 to 255. It defaulted to 225 and I set it to 128, so I'm guessing it's at mode 3 currently.

Share this post


Link to post
Share on other sites
Guest 888

Yeah, looks like the sequential read speed of WD3200KS matches very exactly (in every point) to that WD3200JD/JB has shown at all known 7 testers (including here in SR). Sure these drives are similar, at least it confirms that their platters are just the same.

But yes, the seek time looks awesome slow if to compare with WD3200JD or WD2500KS. It's with AAM enabled but at the same time I think the WD3200JD has also AAM enabled by default (and at testing) because of its seeks are really silent and softly sounding. So this WD3200KS's slow seek time remains a unsolved question yet.

It would be the best way when someone (SR?) who tested the 3200JD or JB would put the 3200KS into the same testbed and compare then...

Share this post


Link to post
Share on other sites
Guest 888

And it is different now!

With AAM disabled it showed up pretty normal access times in your tests, just by 3.2-3.4 ms better than with AAM enabled! Logical, in principle.

Now I finally digged here and there on internet and picked up all known test results for 3200JD, JB, SD:

Random Access Time - READ - Usual Mode 0 - Average:

- WD specs = 13.1 ms

- TH test#1 = 13.3 ms JB

- TH test#2 = 13.2 ms JB

- TH test#3 = 13.4 ms JD

- SR test#1 = 13.0 ms JD

- SR test#2 = 13.0 ms JD

- HA test#1 = 12.7 ms JB

- DL test#1 = 12.9 ms JB

- DL test#2 = 12.7 ms JD

- DL test#3 = 13.6 ms JB

- DL test#4 = 13.4 ms JD

- XYZ test#1 = 13.4 ms SD

- XYZ test#2 = 13.5 ms SD

Random Access Time - READ - Quiet Mode 3 - Average:

- DL test#1 = 16.7 ms JB

- DL test#2 = 16.4 ms JD

- DL test#3 = 17.4 ms JB

- DL test#4 = 17.1 ms JD

So the results are pretty same (differences no more 0.5 ms) as they are in your tests!!!

But the question is, the question remains - is it really so that for 3200JB and 3200JD the default mode is AAM disabled but for 3200KS is AAM enabled??? If it is so then I can't understand why JD/JB are so quiet seeking still. As you said when you disabled AAM for KS then it was very loud...

Share this post


Link to post
Share on other sites

I recently bought 4 WD3200KS for a RAID, but I haven't installed them yet. Exactly what utility (links anyone?) do I need to disable AAM and what setting do I use in it? I read the thread and heard people are using a hitachi utility, but I'm not familliar with it since I haven't bought new drives in quite a while.

Share this post


Link to post
Share on other sites

I need to pickup a drive with good mix of storage capacity and speed at a good price point, i was originally looking at the WD4000YR but this drive seems to be pretty close performance wise and $60+ cheaper. The WD3200KS also has the 5 year warranty if i read correctly and is quieter than the WD4000YR.

Any thoughts on the one vs the other?

Share this post


Link to post
Share on other sites

Here's three links to Hitachi Feature Tool (version 1.99):

http://www.hitachigst.com/hdd/support/down.../ftool_v199.exe

(Windows boot diskette creator)

http://www.hitachigst.com/hdd/support/down...199_install.img

(Boot diskette image for non-Windows systems)

http://www.hitachigst.com/hdd/support/down...s/ftool_199.iso

(CD image with FTOOL)

If you have Windows and a floppy disk drive, click the first link. It's the easies way.

It works both on IBM/Hitachi drives and non-IBM/Hitachi drives. Of course Advanced Power Management (APM) cannot be configured on non-IBM/Hitachi drives simply because those drives doesn't have that feature (it impossible to implement APM without load/unload technology and others use it only in 2.5" and smaller form factor drives).

Seagates AAM setting cannot be configured with FTOOL. It's not a bug in the software, it's simply how Seagate wanted it to be. (I still don't get WHY they don't allow users to reconfigure it. It makes no sense.)

_______________

I have also tried SilentDrive. It a software that allows to change AAM in Windows, no boot diskette required, no reboot required, no installation required (just run the EXE)...

http://www.rt-sw.de/freeware/silentdrive.zip

I'm having some problems configuring SATA drives (mobo integrated controller, SIS Raid) with it, but it can set the AAM mode for PATA drives.

Hitachi FTOOL is more compatible with SATA drives.

Share this post


Link to post
Share on other sites
Guest 888
Do you have to deduct 4.2ms from that 17ms access time to get the "real" random access time or how is that?!

Yes, deducting the latency time 4.2ms you will get the seek time which is usually referred also in official HDD specs (usually 8.5...9.5ms for AAM disabled mode)

Share this post


Link to post
Share on other sites

I'm also interested in how this HD performs (WD3200KS) but there's no damn info or reviews on it as of yet ! Well none that I can find. I had a quick squiz through this thread but dind't really understand most of it :) forgive my ignorance. So for those of you that own the drive do you recommend it? and how do you reckon it would match up against say Seagate's 300GB 7200.9 drive?

Cheers

Share this post


Link to post
Share on other sites
I'm also interested in how this HD performs (WD3200KS) but there's no damn info or reviews on it as of yet ! Well none that I can find. I had a quick squiz through this thread but dind't really understand most of it :) forgive my ignorance. So for those of you that own the drive do you recommend it? and how do you reckon it would match up against say Seagate's 300GB 7200.9 drive?

Cheers

Well my 4 are doing good. I get around 13ms latency and they're really fast in terms of bandwidth. With 4 of them in a 128K stripe I'm seeing about 275MB/sec uncached and 780MB/sec cached reads with hdspeed. I'm not sure how it compares to the 7200.9.

Are there any NCQ-specific benches out there?

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now