Overclock.net banner
Cancel
Create thread New Forums

Western Digital 1TB WD1003FZEX poor access time

Jump to Latest
44K views 40 replies 11 participants last post by  Azerty38  
#1 ·
Well, since my WD1002FAEX is dying I decided to go with the new family of WD's Blacks. So I got the WD1003FZEX and ran some tests, but something seems wrong about the access times, I don't know, I'm not an expert here.

So can anyone tell me how does it look:

HD Tune Pro tests

***
Normal read bench

WD1003FZEX:

WD1002FAEX:

***
Random Access test

WD1003FZEX:

WD1002FAEX:

***
Extra tests

WD1003FZEX:

WD1002FAEX:

***
File benchmark (for some reason I couldn't run this test on the WD1002FAEX)

WD1003FZEX:

***
Crystal Disk Mark

WD1003FZEX:
 
Save
Reply Quote Rep+ Rep+
#2 ·
I was searching info yesterday and apparently for this type of drive, the average access time should be 12ms.

So 16ms is pretty bad I guess.
 
Save
Reply Quote Rep+ Rep+
#3 ·
Are you in AHCI mode, and have you installed all relevant drivers? Some of these new drives perform really bad if you're not in AHCI/RAID mode, or forget to install Intel RST or AMD AHCI drivers. Sequential performance isn't usually affected much, but access times always take a hit.

You may also want to check where you've got it mounted. I've seen cases come with slightly defective mounting racks that transfer more vibrations to certain drives. I could increase any drive's access times by 0.8-1.2ms just by moving it to a certain slot in my case.

If neither of those fixes it, then perhaps these new drives are just optimized for low noise levels or something? (slower seeking) That's not what I'd want to see from a Black drive...
redface.gif
that's fine for Blue, but not Black.
 
#4 ·
Quote:
Originally Posted by Kramy View Post

Are you in AHCI mode, and have you installed all relevant drivers? Some of these new drives perform really bad if you're not in AHCI/RAID mode, or forget to install Intel RST or AMD AHCI drivers. Sequential performance isn't usually affected much, but access times always take a hit.

You may also want to check where you've got it mounted. I've seen cases come with slightly defective mounting racks that transfer more vibrations to certain drives. I could increase any drive's access times by 0.8-1.2ms just by moving it to a certain slot in my case.

If neither of those fixes it, then perhaps these new drives are just optimized for low noise levels or something? (slower seeking) That's not what I'd want to see from a Black drive...
redface.gif
that's fine for Blue, but not Black.
Click to expand...
Yes, the mobo is set to AHCI by default.

I actually realized later that I didn't install Intel RST, so I installed it but didn't make a difference.

Then I tried running a test booting with Mint on a USB drive, and the HDD throws 15.49ms all the time.

So I put it on my old AMD PC running Windows 7, same thing, 16ms, I ran Mint from that same PC, same thing, 15ms.

I actually put it on a flat surface when I hooked it up to the AMD PC.

So yeah after a talk with Western Digital and a DLGDIAG extended test showing no errors, I ended up returning the drive. My replacement gets here today, so I'm crossing my fingers.

One thing I did notice is that checking the features of this drive, AAM is not supported, which seems really weird. The drive is super quiet.

What I'm thinking is that the HDD is exchanging performance for less noise, and there's a bug with the firmware or something that's not letting me tweak the AAM parameters and shows it as a not available feature.
 
Save
Reply Quote Rep+ Rep+
#5 ·
Quote:
Originally Posted by xutnubu View Post

One thing I did notice is that checking the features of this drive, AAM is not supported, which seems really weird. The drive is super quiet.

What I'm thinking is that the HDD is exchanging performance for less noise, and there's a bug with the firmware or something that's not letting me tweak the AAM parameters and shows it as a not available feature.
Click to expand...
It may not be a "bug" by their definition - it may be a "feature" - but the customer service guys wouldn't know... that's the kind of thing an engineer might.

As platter densities increase, it gets harder and harder to seek to the correct track. WD prides their Black line on delivering a consistent great experience. It's possible they just couldn't get every drive down to 12-13ms at this density, so they went for noise levels instead. Time will fix it, as the manufacturing process matures - but who wants to wait?
tongue.gif
Report back on whether the new one is any better. If not, I'd write it down as a WD1003FZEX "feature". No AAM, super quiet, fairly high seek times, similar performance to older slower Blacks. (only sequential performance is higher... which affects file copies.
redface.gif
Woohoo... )

Also, considering the number of posts I've seen where people RMA'd because their Black was noisy, I suspect the number of RMAs due to high seek times will be lower than that. This is annoying for enthusiasts that buy the drives not caring about the noise levels, but we're probably the minority of their customers. I know a lot of shops/users/businesses/etc. don't even run extended SMART tests on drives before putting them into use, so they'd never catch a thing like this.
 
#6 ·
Got the replacement, same thing.

I'm frustrated now, I don't know if this is actually how the drive works or what.
 
Save
Reply Quote Rep+ Rep+
#7 ·
Do you think it has something to do with the drive using only one platter?

I just read some Newegg reviews and I found one on the 1Tb model and the guy reported an access time of 16ms.
 
Save
Reply Quote Rep+ Rep+
#8 ·
Quote:
Originally Posted by xutnubu View Post

Do you think it has something to do with the drive using only one platter?

I just read some Newegg reviews and I found one on the 1Tb model and the guy reported an access time of 16ms.
Click to expand...
No - it has something to do with how they configured the firmware. For low noise levels rather than fast seek times.
redface.gif


This doesn't really help you, but I just picked up a WD4001FAEX on Boxing Week - it's a 5x800GB platter drive, configured for fast access times rather than low noise levels.

There probably wasn't any point going for low-noise with a 5-platter design.
tongue.gif

zjlj.png

61lc.png


Whole drive, it's on-par with your WD1002FAEX.

Only 1TB tested, it's obviously far faster. (The law of short-stroking!)

p6ox.png

8mtb.png
 
#10 ·
Quote:
Originally Posted by SkyNetSTI View Post

Let me join to the thread...
Just picked up WD1003FZEX and my test showing even worse results...
What is going on?


Click to expand...
Seems to be common for this drive.

I contacted them (WD) several times and I never got a definitive answer, on top of that their support is slow as...

Your throughput results are like that because I think your drive is half full. Ideally, you should test it when empty. I tested mine on a fresh install of Windows 8.1, so it was only using 20 something GBs.

Yours seems to throw the same access time numbers as mine, so about 16ms.

I believe there's also an overhead introduced by Windows, you can run a benchmark using something like Mint (usb bootable), I always get around 154MB/s and 15.49ms using that, in contrast, HD Tune in Windows always throws 143MB/s and 16.6ms or worse.
 
Save
Reply Quote Rep+ Rep+
G
#11 ·
Hi. I am also seeing these much higher than typical (for WD Black drives) Access Times on my 1003FZEX. However, I do not know whether or not this is actually a problem. From what I can tell so far, the sequential as well as the random read/write results I'm seeing on CrystalDiskMark are quite good regardless of the Access Time being upwards of 16ms.

If I understand correctly, the Access Time is largely going to affect high numbers of small random reads and writes. But according to CrystalDiskMark, for HDDs the random read and write speeds of my WD1003FZEX are fairly good and at least on par with what the 12ms drives (like the WD4003FZEX) are providing.

Perhaps regardless of the higher Access Times the throughput for random read/write is just as good as the lower Access Time drives because the 1003 has a higher density single 1TB platter? And the bits are thus being read faster once the head gets to where it needs to read them? I'm guessing.

Here are my HD Tune results:



Also note the strange outlier access time points around 2% and 78% of the drive (at 28ms and 39ms respectively). Those points are always there and I don't know what that means. I have another forum post inquiring about it.

Now here are my CrystalDiskMark results.



And finally here is a CrystalDiskMark result I found on the internet for the 4TB version. The WD4003FZEX. Notice that my results for the 1003FZEX are on par with this drive.



So... are these values actually okay for my 16ms Access Time WD1003FZEX ? Why are my values essentially the same as for the 4TB version of the drive which uses 5 - 800GB platters, as opposed to my 1 - 1000GB platter.

Xutnubu, did you ever find out if these Access Times are normal for these drives, and we don't simply have defective ones? From what I've seen so far (though there is currently little information out there), all these 1003FZEX drives are in the 16ms range.

Kramy, you seem to be the guru around these parts, what are your thoughts on my benchmark results? Is this a pretty good drive after all, regardless of the seemingly high Access Time?

Thanks.
 
G
#12 ·
By-the-way,

I should note that I'm using an old GigaByte EX58-UD3R motherboard, and it only operates at SATA2 interface speeds. Which I think is why my burst speeds are a good deal lower than what others are seeing on these drives. (Can anyone confirm this?)

I am in AHCI mode with the most recent Intel RST drivers I could find that are still compatible with an ICH10R chip set.
 
G
#14 ·
See now I don't get that.

Those results for your Blue drive appear to be better than this recently released 1TB WD Black drive, which is purported to be a good deal better than their previous Black high performance drives. The Blue line being their middle-of-the-road model.

Have you done anything to adjust the performance parameters of your drive. I'm going to have to look around now to see if your results are typical. Blue drives are cheaper than Black and, in addition to the shorter warranty, are supposed to be lower performing than Black (let alone the most recently released Blacks). In fact, other than the higher Access Time, your throughput results in HD Tune are better than even the higher performing 2TB+ newest Black models.

I'd suspect my SATA2 interface, but my results seem typical of other WD1003FZEX drives that are using SATA3 interfaces.

What's your secret?
tongue.gif
 
G
#16 ·
Well I noticed that you were using HD Tune Pro which is version 5.00, and I was using the older HD Tune version 2.55.

I also notice that your screen shot said trial version, which made me realize I could get the trial version of Pro.
thumb.gif


At any rate I downloaded it, and ran the benchmark. The results were basically identical to the non-Pro version. Although I am getting much better Burst Rates with the Pro version. Still not as good as I suspect they could be if I had a SATA3 interface.

Here are the results. Which are seemingly lower than your Blue drive.



Do you have any CrystalDiskMark results for your drive?

I want to see if basically Black is now just paying for a significantly longer warranty rather than increased performance over Blue.
 
G
#18 ·
Hmmm. Yeah.

I don't know how much impact having lots of data on the drive impacts the CrystalDiskMark results.

It would really surprise me if the Blue performance is on par with, or even better than the Blacks.

If that's the case, then it would seem Western Digital has found a way for consumers to purchase an extended warranty for similarly performing drives, and calling them "performance".
devilsmiley.gif
Best Buy must be jealous.
 
#20 ·
Quote:
Originally Posted by Gleno View Post

See now I don't get that.

Those results for your Blue drive appear to be better than this recently released 1TB WD Black drive, which is purported to be a good deal better than their previous Black high performance drives. The Blue line being their middle-of-the-road model.

Have you done anything to adjust the performance parameters of your drive. I'm going to have to look around now to see if your results are typical. Blue drives are cheaper than Black and, in addition to the shorter warranty, are supposed to be lower performing than Black (let alone the most recently released Blacks). In fact, other than the higher Access Time, your throughput results in HD Tune are better than even the higher performing 2TB+ newest Black models.

I'd suspect my SATA2 interface, but my results seem typical of other WD1003FZEX drives that are using SATA3 interfaces.

What's your secret?
tongue.gif
Click to expand...
Quote:
Originally Posted by Gleno View Post

Hmmm. Yeah.

It would really surprise me if the Blue performance is on par with, or even better than the Blacks.

If that's the case, then it would seem Western Digital has found a way for consumers to purchase an extended warranty for similarly performing drives, and calling them "performance".
devilsmiley.gif
Best Buy must be jealous.
Click to expand...
According to Passmark's online benchmarks, the Blue is indeed faster than the Black. The EZEX scores at 1107 while the FZEX scores at 987.
 
G
#21 ·
Quote:
Originally Posted by Techie007 View Post

According to Passmark's online benchmarks, the Blue is indeed faster than the Black. The EZEX scores at 1107 while the FZEX scores at 987.
Click to expand...
Well this situation seemed a bit outrageous to me. I went to check out those PassMark scores, and you snagged the wrong drive for the FZEX. You wanted the WD1003FZEX.

So with the right drive selected, the new 1TB Black does score marginally better than the newest 1TB Blue. But not by very much.

The WD1003FZEX score is 1131. While the WD10EZEX scores 1107.

Average price $88 for the Black and $65 for the Blue. Giving the Blue a much higher Drive Value (that of course does not take into account the warranty differences).

Had I known these drives were so close in performance I might have saved myself $20 and gone Blue. So... is $20 worth an extra 3 years of warranty?
 
G
#22 ·
And presumably you now get some of their enterprise drive features in these new Blacks, like vibration control and corruption protection, whatever those are. No idea if that's just a gimmick or if that is actually meaningful.

If you go to Western Digital's web site, their marketing for the Black drives is A LOT fancier and showy than it is for the Blue. (Maybe that's where the extra $20 goes
biggrinsmiley.gif
)
 
#23 ·
Quote:
Originally Posted by Gleno View Post

Well this situation seemed a bit outrageous to me. I went to check out those PassMark scores, and you snagged the wrong drive for the FZEX. You wanted the WD1003FZEX.
Click to expand...
Thank you for the observation. However, the only difference I see between those two entries is the presence of a space between "10" and "03" ("WDC WD1003FZEX" vs. "WDC WD10 03FZEX"). Looking even closer at the individual PassMark pages for those drives (here and here), the "sub-number" listed in "Other names" (-00MK2A0) is the same. I honestly think that both entries are referring to the same drive.

My question now is, "Why the space, and why the lower score on the drives with the space?" A Google search for "WD10 03FZEX" didn't turn up anything relevant (as if the part number was invalid), while a search for the same number without the space between "10" and "03" brings a pile of results. I wonder if WD had two different factories producing (or making parts for) this drive, or two "secret" revisions of the drive. Just out of curiosity, could those reading this thread post what model number appears for their Black/FZEX drive in HwInfo? For the Blue/EZEX, I'm getting "WDC WD10EZEX-00ZF5A0", and I've gotten peak sequential read speeds over 200 MB/s from it.
 
G
#24 ·
Hmmm. Curious.

Anyhow, the version of my drive is listed as WD1003FZEX-00MK2A0.

My peak reads are just below 190 MB/s, so basically on par with what your Blue is getting.

I'm beginning to think that, at least for the 1TB single platter, that the Blue and Black are essentially the same, and you're paying a minimum $20 extra for 3 years added warranty.

Of course these are synthetic benchmarks, so who knows. There are presumably other advantages to the way the Black functions, dual 64 bit processors, special algorithms, etc, that might make real world performance and reliability better than the Blues. I don't know if WD claims whether the Blue drives have these features they are touting on their Black drives or not.

But based on the synthetic benchmarking the drives appear to perform identically.
 
#25 ·
Quote:
Originally Posted by Gleno View Post

Hi. I am also seeing these much higher than typical (for WD Black drives) Access Times on my 1003FZEX. However, I do not know whether or not this is actually a problem. From what I can tell so far, the sequential as well as the random read/write results I'm seeing on CrystalDiskMark are quite good regardless of the Access Time being upwards of 16ms.

If I understand correctly, the Access Time is largely going to affect high numbers of small random reads and writes. But according to CrystalDiskMark, for HDDs the random read and write speeds of my WD1003FZEX are fairly good and at least on par with what the 12ms drives (like the WD4003FZEX) are providing.

Perhaps regardless of the higher Access Times the throughput for random read/write is just as good as the lower Access Time drives because the 1003 has a higher density single 1TB platter? And the bits are thus being read faster once the head gets to where it needs to read them? I'm guessing.

Here are my HD Tune results:



Also note the strange outlier access time points around 2% and 78% of the drive (at 28ms and 39ms respectively). Those points are always there and I don't know what that means. I have another forum post inquiring about it.

Now here are my CrystalDiskMark results.



And finally here is a CrystalDiskMark result I found on the internet for the 4TB version. The WD4003FZEX. Notice that my results for the 1003FZEX are on par with this drive.



So... are these values actually okay for my 16ms Access Time WD1003FZEX ? Why are my values essentially the same as for the 4TB version of the drive which uses 5 - 800GB platters, as opposed to my 1 - 1000GB platter.

Xutnubu, did you ever find out if these Access Times are normal for these drives, and we don't simply have defective ones? From what I've seen so far (though there is currently little information out there), all these 1003FZEX drives are in the 16ms range.

Kramy, you seem to be the guru around these parts, what are your thoughts on my benchmark results? Is this a pretty good drive after all, regardless of the seemingly high Access Time?

Thanks.
Click to expand...
I asked their support several times, they sent me to the engineering department and I never got an answer.

My guess is that because this drive uses 1 platter you can't simply see the access time and say that it sucks, perhaps there's some dark magic going on that can't be shown in numbers that puts the drive in a "not too bad territory" despite the high access time figure.

But yeah, if you don't mind the warranty and couple of extra features, I think it's better to just go with the Blue.

BTW I asked the support agent if this drive has VCT (vibration control) StableTrac because if you go to the official website, under features, it is specified that only 2TB models and above have the feature, but the guy told me that the 1003FZEX has it as well. Then I asked him why that statement was contradictory to what you read on the official site, and he answered again that the 1TB FZEX has VCT StableTrac.

Their support is a joke.
 
Save
Reply Quote Rep+ Rep+
G
#26 ·
Yes the 1TB do also have the VCT. You are thinking of StableTrac, which is only on the 2TB and above and is only really relevant to drives which have more than 1 platter, and require additional stabilization.

The VCT, high resolution controller (HRC), 64 bit dual processor, dynamic caching algorithms, (which are things I'm not sure if the Blue drives have) should presumably make these drives higher performing. If that is indeed the case, then I can only presume that it won't be evident in synthetic benchmark tests against their Blue drives, and will only be obvious in real world use applications, like gaming.

I don't know that the 16ms access times are a problem. The higher density platters apparently make up for these higher access times, because once it reaches the position it needs to read from it reads faster from a higher density platter. (So they read a small file in the same amount of total time as a lower Access Time drive). This may be why when comparing performance results between the 1TB drives (using a 1TB platter) and the higher TB drives with 12ms Access time (using 800GB platters) they are pretty much on par with each other, and they are all on the high end of drive performance from what I can tell. I'm no expert.

I would like someone who has more understanding of these things to chime in and reassure us 1TB Black drive purchasers that we got our money's worth by spending a fair amount more on the Black drive instead of the Blue, and that these high Access Times are nothing to be concerned about for this high density single platter drive.
 
Overclock.net
posts
28M
members
562K
Since
2004
The premier forum for overclocking experts and enthusiasts. Discuss hardware optimization, custom builds, benchmarking, cooling solutions, and pushing the boundaries of computing performance. From beginner guides to extreme overclocking, join our technical community to master system tuning.
Full Forum Listing
Explore Our Forums
Intel CPUs Hardware News NVIDIA PC Gaming Water Cooling