Exploring WD's Advanced Format HD Technology - HotHardware

Exploring WD's Advanced Format HD Technology

10 thumbs up

Worldbench 6b2

We focused on three tests out of the Worldbench 6 suite: Adobe Photoshop CS2, Nero 7 Ultra, and Winzip.



 



 



Once aligned, the Caviar Green does a reasonable job of hanging with the Caviar Black considering that the latter is a higher-end drive. If left unaligned, the Caviar Green's performance craters; it takes twice as long to run through the Nero 7 Ultra test, is 3x slower in Adobe Photoshop CS2, and takes more than 3.5x as long with the Winzip benchmark. This was not an isolated incident, as you'll see from our File Copy and Sandra benchmarks.



 


FC-Test is a benchmark that creates and copies files from Point A to B and records the total time. We copied two types of files from the test—MP3's (271 files, 990MB, average file size of 3.65MB) and programs (8504 files, 1380MB, average file size of 162K.) While the MP3 write times take a serious hit due to misalignment, it's the small-file program writes that again kick the legs out from under the Caviar Green. SiSoft Sandra 2010's file system benchmark further confirmed the shape of our analysis, and as expected, random write performance between Aligned Green and Black was almost identical.

Finally, we switched to 64-bit Windows 7 and ran PCMark Vantage on both drives.


Ironically, the performance gap between the Black and the Green is larger here than in almost any of our other tests. The Caviar Black is slightly more expensive (currently listing for $99.99 at NewEgg compared to $84.99 for the WD10EARS. There's also a newer Caviar Black on the market with 64MB of cache and support for SATA 6G. This drive is supposedly faster than our current Caviar Black even when using SATA II, and is retailing for $119.99.

Western Digital has done a good job of providing the necessary tools to ensure that a new Advanced Format drive works under Windows XP; the benchmark results we've shown you are meant to underscore the need to use the utility and/or jumper setting as the company suggests if you plan to use Windows XP. This last scenario seems likely, given that Windows XP is still the most popular OS according to Steam, with 43.81 percent of the market. Since read speeds are much less affected by the misalignment, anyone who buys an Advanced Format drive and plugs it in without reading the instructions could end up bewildered when the drive performs perfectly in some ways and badly in others.

As for Advanced Format, there's not currently much to say. Western Digital believes the technology will prove useful in the future and it's true that after thirty years, the 512 byte sector standard was creaking with age. WD is not the only company planning to implement Advanced Format, but it may be a year or so before we can access the true impact of the new standard. The only thing to be careful of when comparing AF and non-AF drives is average transfer rates. This is not currently an issue, and hopefully it never will be, but a manufacturer that uses AF to avoid using the inner tracks of the drive platters can claim higher average transfer rates without actually increasing performance at all. At $84.99 for 1TB, the WD10EARS is a solid value and we'd recommend picking one up if you want a forward-facing storage solution with solid performance and a low price.
 

Article Index:

1 2 Next
0
+ -

Do you guys think that WD will enable this technology for the Black drives at any time?

0
+ -

Anytime?  Definitely. My guess is that WD went ahead and did the Caviar Green drives first because it was a change that could be easily slipstreamed into the 64MB cache versions of the disk. 

The Black series just got a refresh; the new 64MB cache drives are a fair bit faster than the old 32MB and they offer SATA 6G. I'm betting we'll see Black AF drives the next time WD refreshes the product series.

0
+ -

It'd be interesting to see how a SATA 6G Caviar Black drive with this new feature and the 64 MB cache fared against a SATA 3G SSD. If the results were close, it could mean new life for mechanical HDs (though I'm sure a 6G SSD would beat the tar out of them).

0
+ -

At least from early benchmarks I've seen and based on current hard drive technology, I don't think this improvement will be anything too far (in performance) from what we see today in traditional platter drives. Aren't the Caviar Black drives not even reaching anywhere close to the 3 GB/s that the "older" SATA offers? I don't think there will be much difference from a 3 GB/s to a 6 GB/s platter drive. Maybe, though I still don't think so, there might be some difference when in RAID.

I think any real improvement in the platter drive would be to make a huge cache (in GB), which is basically like that hybrid design that Silverstone came up with. Other than that, I think SSDs are the way to go in terms of pure performance.

From my understanding, I think the main reason for this Advance Format is to make it easier to produce higher capacity hard drives. Correct me if I'm wrong.

0
+ -

Clem,

I think you're confused on a few points here--let me see if I can help.

Performance: Neither SSDs nor HDDs can keep 6Gbps of bandwidth fully saturated right now. The difference is that mechanical hard drives will likely never be able to pull this off. The gap between SSD and HDD performance is variable depending on what sorts of benchmarks one runs (and SSDs don't always have a huge advantage), HDD's leverage their tremendously lower cost-per-GB vs. SSDs, not raw performance.

#2. SSDs generally are the way to go for pure performance, but unless you've got a hugely fat wallet, price inevitably comes into play. This is where things currently get sticky--there are a variety of drives and controllers on the market with a huge variety of performance. If you can afford to spring for an enterprise-class, $800 SSD (or two in RAID), yes, it'll blow an equivalent $1600 HDD performance setup out of the water and it'll draw a lot less power. If, on the other hand, you're trying to balance price and storage and don't want to spend more than $150, a new Caviar Black [insert manufacturer of your choice here] will be the better option.

There is a definite diminishing marginal return that makes "building a huge cache" a bad idea past a certain point. Remember, the entire point of a cache is to provide the system with a low-latency/high-bandwidth storage area in which only highly relevant data is stored.

The more cache in a system, the longer it takes to search it and the more unwieldy it can be to maintain. The more time the drive has to spend keeping the cache updated and filled, the greater the chance that the end-user will actually notice performance degradation.

Finally: There is no one "main point" to AF. It does or enables a series of options; the larger disk drives is just the simplest to understand.

0
+ -

Normal 0 false false false EN-US X-NONE X-NONE MicrosoftInternetExplorer4

I don't think Clem is confused, his posts are generally on different things than what you are talking about in some points. Also the cache depends largely on how it is handled. Yes SSD's are the way to go for pure performance to a point. However; a system that is focused on performance alone, will also to some points not be the best, because everything is focused on performance alone, and therefore usability is sacrificed. I personally am currently running 2 WB black 32Mb cache drives. However right now I don't have them in a RAID configuration because to the greatest point it is not really necessary.

I am currently thinking about grabbing a SSD for one thing and one thing only which would be main app's, the largest would of course be the OS, and a dual system image backup to the SSD, and to the converted RAID0 with 2 WD Caviar Black 750GB drives. Of course I have not implemented it yet I think this would be the best combination. I would have speed from the SSD platform; I would have space from the RAID configuration, and would also have performance with it being Raid 0.

As for your comments on a HDD catching up to and SSD you are 100% correct that will never happen. Also on top of this SSD's are new devices, and will therefore be upgraded for quite some time, this is very valid in this discussion because we are basically talking about the first large storage format HDD in a commercial market Vs. a brand new one, and the old one is as in this article still being upgraded.

On the more cache thin and the implementation thereof someone mentioned one thing in this conversation which I think is very valid. The combination of an SSD for speed and cache duties and a modified HDD into a singular device seems very promising. This would and could be very nice on all ends, but I think the controlling scheme as well as software firmware etc needs a good bit of work to find the most efficient, and speedy implementation scheme. This in many ways is much like I am saying I may do on my PC, with the RAID 0/SSD setup where the most active programs and software will be on the SSD for performance, the rest on the HDD for its capacity and relative speed in a RAID 0 2 x  750Gb 7200 drive scheme.

 


0
+ -

Ah, thanks, Joel. Wasn't what I was asking, but thanks for the reply.

0
+ -

Bout time, now if they can just give us adjustable speed up to 10-12K! Without killing us for the comparable size.

0
+ -

Joel, do you know if WD have put something in their WDAlign tool that prevents it from being run on non-WD drives?  If not, then this tool could be a godsend for people with mis-aligned SSDs.  Aligning an SSD for XP use can sometimes be quite awkward, so if this could do the trick I'd be delirious.

0
+ -

"A WD10EARS and a WD10EADS have exactly the same unformatted capacity and Windows reports both drives offer 931GB of storage space."

 

That does not debunk the myth as I understood it. I have read that the "usable space" is affected. Meaning that if both drives are filled with files. The WD10EARS will be able to hold more files/more data. That is because it is more efficient at storing information. Am I correct?

1 2 Next
Login or Register to Comment
Post a Comment
Username:   Password: