by Alexey Volkov , Nikita Nikolaichev
03/16/2007 | 03:32 PM
After we covered hard disk drives with the highest storage capacity available today in our 500GB HDD Shootout: Seagate Barracuda 7200.10 and Others!, it sounds reasonable now to turn our attention to the other end of the capacity scale (showing some volatility typical of creative persons).
So what do we find there? We are still offered a handful of 40GB models and a rather large number of 80GB models. But is it worth the trouble reviewing such HDDs? Besides creative volatility, the reviewer must have an analytical mind in the first place.
So, let’s analyze the price/capacity ratio of modern HDDs with a SATA interface (we took the price list of a neighboring computer shop for that purpose).
As you see, this ratio becomes reasonable with 160GB drives. HDDs of smaller capacities may only be interesting for users who want to save some money (even though as little as a few dollars) and don’t need a large storage capacity. When you come to think of it, 40 gigabytes of hard disk storage is still quite enough for an office computer, but is too little for a home PC with all those games, music and video.
In this roundupwe’ll test 160GB HDDs with the SATA interface only (because if we took all 160GB drives irrespective of the interface, the diagrams just wouldn’t fit in the screen). Some of them are modern and some are not. You won’t find the results of the newest drives from WD and Samsung here – we’ll check them out in the second part of the review where we are going to compare all single-platter drives, perpendicular and otherwise. There is a lot of exciting matter in this review anyway, as you’ll see soon.
Now it’s time to get acquainted with the HDDs to be tested.
Hitachi is represented here by two models from the Deskstar T7K250 and 7K160 series. The older of them, the HDT722516DLA380 , is the junior model in the super-short Deskstar T7K250 series which includes only four models with two different capacities (160 and 250GB) and two interfaces (ATA and SATA).
The case of the HDD is typical of Hitachi’s multi-platter drives. The drive doesn’t have any special traits. It is an ordinary “first-wave” SATA-300 device with an 8MB buffer, support for NCQ and delayed start, etc. Being a rather aged product, the HDD is based on two platters of which it uses only three surfaces, though. Thus, the capacity of one platter is about 105GB.
The HDS721616PLA380 model from the Deskstar 7K160 series is quite a different story. It can be noticed with a naked eye:
The HDD case is made as light as possible – it offers room for one platter only. As a matter of fact, it does contain only one platter J. The platter is 160GB large and this capacity is achieved using the good old non-perpendicular recording method.
Depending on the HDD capacity, either one or two surfaces of this platter are employed. Easy to count up, the 7K160 series consists of only four models, too.
The capabilities of the HDD’s electronics are the same as those of the T7K250 series like an 8MB cache and NCQ. Thermal Fly Height Control technology is declared (when the air temperature inside the HDD case changes, it is necessary to correct the aerodynamic properties of the actuator’s tip to maintain the required height of the head above the disk surface).
The specifications of the Hitachi drives are listed in the following table:
Following in alphabetical order, Maxtor goes next.
We’ve got two HDDs from Maxtor for this review. The first is a refresh version of the DiamondMax 10 . The second is the company’s swan song, the DiamondMax 17 . Yes, there had appeared two more HDD models under the Maxtor brand – DiamondMax 20 and DiamondMax 21 – before the company was devoured by Seagate, but those were Seagate’s HDDs from the Barracuda 7200.9 and 7200.10 series with a Maxtor label and slightly revised firmware. Let’s start with the DiamondMax 10:
It looks like a regular DiamondMax 10, but take note of the second digit in the drive’s serial number “V300GJ7G”. This number indicates three read/write heads whereas the original 160GB DiamondMax 10 has four heads!
The second difference from the original DiamondMax 10 is new SATA-300 electronics. Unfortunately, the drive’s cache buffer was not increased to 16MB.
The DiamondMax 17 resembles the previous model, except for the bottom view:
The PCB is turned with the chips facing inward and is also smaller (which is an important thing for some hardware reviewers). However, the electronics is no different functionally: SATA-300 and 8 megabytes of cache memory. If you take a look at the drive’s serial number, you’ll see its main difference from the previous model. The DiamondMax 17 is a single-platter drive!
Having one platter and one head less is important for this class of HDDs as it helps reduce the prime cost considerably. Well, Maxtor was too late, anyway…
Samsung is represented in this review by two devices: the SP1614C model from the SpinPoint P80 series and the HD160JJ from the P80 SD series. The former is not just old, it’s very old. It has twice taken part in our tests already.
The second drive from Samsung is a more interesting thing. It is a refresh version of the old P80 series: old bones, new electronics…
Well, new technologies (SATA-300, NCQ) must be supported, even if just for show. Or you’ll have to be selling your drives at much cheaper prices than the competitors, which is not a good market policy :).
Seagate has had a traditionally broad presence in our tests. It is not that we are biased towards this brand (we are equidistant from all the manufacturers), but Seagate’s product range has been always wider than that of any of its competitors in the last few years. Particularly, there are so many 160GB drives in the Barracuda 7200.9 series that we just couldn’t count up all the varieties.
It’s like the term “HDD series” doesn’t mean something constant anymore. HDDs are being updated and modernized on the go and continue to be produced under the old name instead of getting a new one. We’ve got used to such updates with CPUs (steppings, revisions), but it’s somewhat confusing yet to have the same thing with hard disk drives.
So, let’s sort this all out. It is easy with the first drive. It is a Barracuda 7200.7 with native SATA-150 electronics (i.e. without additional PHY-interface chips and with support for NCQ).
The drive is old, but it is going to show you some yet. Seagate wasn’t silly in choosing it to showcase the advantages of NCQ.
The next HDD, the ST3160812AS model, belongs to the Barracuda 7200.9 series (by the way, don’t think we’ve forgot the 7200.8 generation – it just doesn’t include models with capacities smaller than 200GB).
So, the case of this drive is somewhat different from the case of the big brother from the Barracuda 7200.9 series. It is thinner and, obviously, lighter. There is a dent in the top panel opposite the actuator (to fix it in place? to dissipate heat?)
In our previous review we found out that Seagate had abandoned its “one design for all the models of a series” principle with Barracuda 7200.9. And now it seems Seagate had gone even further than we suspected.
Here is yet another 160GB drive from the Barracuda 7200.9 series!
Note that the case has become even thinner and the PCB has been turned around so its chips now face inwards. Moreover, it has become smaller, which seems to be a tendency. Why such a dramatic update of junior models in the series? Just to save some aluminum and textolite?
There is a supposition (not confirmed officially yet) that Seagate’s desire to transfer HDDs from the junior branch of the Barracuda 7200.9 series to perpendicular recording was the reason for this update. Seagate’s production volumes are huge, and it is very desirable to unify the main subunits of hard disk drives the company is turning out.
The next drive is a copy of the previous one, except for the amount of cache memory. The ST3160211AS has only 2 megabytes of memory, which is very little by today’s measures.
Western Digital has only one drive in this roundup.
The case has a rather gloomy color – the manufacturers sometimes use such tricks as an unusual color to attract the potential customer’s attention. We hope Western Digital doesn’t play cheap tricks to us. Perhaps the company just shows its mourning for its late competitors J.
The WD1600JS belongs to the Caviar SE series and thus has an 8MB cache buffer. Like one of the Hitachis, the model is rather old and required three surfaces from two platters to achieve the capacity of 160GB. There are dual-headed models of the same capacity available now which can be distinguished by two extra letters in the model name - WD1600AAJS . The two magic letters AA may do miracles, but we’re running too far ahead…
Let’s get back to the specs:
That’s all the HDDs we are going to test today.
The IOMeter tests are performed on non-formatted disks. For WinBench 99 the hard disks are formatted in NTFS for their full capacity and the data-transfer graph is recorded. Then a 32GB partition is created on them and formatted in NTFS and then in FAT32 to perform the main test program. To reduce measurement errors, we run the test seven times and calculate the average.
For the File Copy Test two 32GB partitions are created on the disks, formatted in NTFS and then in FAT32. The partitions are created at the beginning of the disk address space.
PCMark04 and PCMark05 run on the full disk capacity formatted in NTFS. These tests are performed ten times each and the results are averaged.
We can’t use the Database and Sequential Read and Write patterns to compare HDDs since these patterns yield a colossal amount of data and putting all this data into one table or diagram would be unreadable. So, if you want to see how the HDDs perform in these patterns, follow this link for tables and diagrams.
We’ll start our comparisons with patterns that simulate the load on the disk subsystem of a server.
Red marks the worst results and blue marks the best ones. The Rating column contains the average results of the HDDs which we used to build the following diagram for easier comparison:
Surprisingly, it is the Seagate ST3160827AS – one of the oldest devices in this review – that occupies the first position and with a rather big lead over the others. After it the HDDs have lined up with minimum gaps. Funny enough, Seagate’s products take both first and last places here.
There are two winners in this test, one of which is the same as in the previous one. And both winners are three-headed, i.e. old models. Obviously, the manufacturers don’t think it necessary to optimize their new HDDs for this type of load.
The results of this pattern are strongly influenced by the average access time as well as deferred writing algorithms. If we consider the results of the synthetic Database pattern, we’ll see that the two leaders in the Workstation pattern are superior in these very parameters. The Seagate ST3160827AS has an excellent access time whereas the WD1600JS works superbly when there are a lot of writes to be performed.
Let’s now see what happens if the address space of the test is limited.
So, both drives have gone even farther away from their pursuers.
Note the sluggish operation of Maxtor’s HDDs which are both very poor in this test. On the other hand, Maxtor’s HDDs have always been good at processing paralleled requests which we are going to check out in the next test.
Follow this link to see detailed tables containing results of this test. We’ll discuss diagrams that compare the drives at a requests queue depth of 1.
Seagate’s drives with perpendicular recording take all of the podium when there is only one thread to be processed. When data is being read in one thread only, this test transforms into a sequential reading test and the drives line up according to their areal density. Somewhat boring, eh? Let’s make it hotter now by increasing the number of simultaneously running threads!
With several threads, the HDD cannot win through sheer physical force alone. It is important that the drive’s firmware could recognize the threaded nature of load and could adapt the read ahead mechanism accordingly.
That’s it: Seagate’s programmers just didn’t think about such loads. As a result, the company’s drives almost refuse to work, delivering indecently low read speed. Maxtor’s drives don’t have such a high sequential read speed as the last generation of Seagate drives, but they use what they have in a most efficient way. The Maxtors’ speeds of reading in two threads do not differ much from their sequential read speeds.
The Hitachi HDS721616PLA30 is the closest drive behind the Maxtors, but its speed is lower more than by half. Funnily, Samsung shows some degradation in terms of firmware: the older SP1614C is much faster than the newer HD160JJ.
Nothing changes when it comes to reading three and four threads. The Samsung SP1614C and the WD1600JS drives can be marked out. They are slower than the Maxtors, but are not as poor as the others.
Now let’s see what we have at writing.
Strangely enough, the Hitachi HDS721616PLA30 is the best one when writing one thread, although it doesn’t use the new perpendicular recording technology as its pursuers, three models from Seagate’s last generation of HDDs, do.
The Maxtor drives take two top places when we add a second thread, but their opponents don’t look as hopeless here as they did in the multi-threaded reading test.
The single loser of this test, the Seagate ST3160211AS, has a valid excuse. It is the only HDD in this review to have only 2 megabytes of cache memory.
The picture doesn’t change much as we add a third and then a fourth thread. The Maxtor drives are still ahead, while the Samsung HD1600JJ and the Seagate ST3160827AS are contending for third place. Note also how the Seagate ST3160827, the oldest HDD model in this review, leaves most of the newer HDDs behind in this test.
We use WinBench 99 to check hard disk drives in desktop PC mode. To do this, we format the disk in NTFS and FAT32 with the OS’s formatting tools.
Here are the linear reading graphs of the HDDs:
Let’s first compare the access time of the HDDs.
Low recording density has a positive effect on the average access time (through the use of more read/write heads and, accordingly, through the reduction of the work zone of the platters). That’s why the two rather old HDDs Seagate ST3160827AS and Hitachi HDT722516DLA380 have the best results.
And here’s the sequential read speed of the HDDs at the beginning and end of the logical volume:
Surprisingly, the Hitachi HDS721616PLA360 proves to be the fastest disk again, and it has the highest sequential read speed on both innermost and outermost tracks. It is followed by the three drives from Seagate’s new series, but all of them are inferior not only to the leader, but also to the WD1600JS, on the slowest tracks. Sometimes the required capacity is achieved by using the platter to the full…
Now let’s check the HDDs out in two integral tests: Business Disk Winmark and High-End Disk Winmark. We created 32GB partitions on the disks to do that.
The Hitachi HDS721616PLA360 is in the lead again. Its closest opponents Maxtor 6V160E0 and Seagate ST3160812AS change places depending on the pattern.
And the same test once again, but in the FAT32 file system.
The leader is the same, but the second position in the High-End test is now occupied by the WD1600JS which pushed the Maxtor 6V160E0 down to third place. In the Business Disk Winmark test second place goes to the Hitachi HDT722516DLA380 while the Samsung HD1600JJ is third.
We’ll use FC-Test following to our traditional method. We create two 32GB partitions and format them in NTFS and then in FAT32. A file-set is created in the first partition. This file-set is then read, copied into a folder in the same partition (Copy Near) and finally into a folder in the second partition (Copy Far).
Let’s start with NTFS. Tabled results can be viewed here , and we’ll use diagrams.
The first diagram shows the speed of creating a set of files on the disk.
The Maxtor 6V160E0 copes with the task of creating the Install file-set better than the others. Interestingly, three HDDs from three different companies occupy second, third and fourth places with similar results. There is a ready explanation, though. These HDDs all have similar areal density and the same amount of cache memory. The latter parameter is very important here, by the way. Take note of the difference in speed between Seagate’s last-generation Barracuda 7200.9 drives (the ST3160811AS and ST3160211AS models) that differ in the size of the cache buffer.
Maxtor 6V160E0 appeared the leader again, while the second Maxtor rolled back into the fourth place. The second fastest was Hitachi HDS721616PLA380, it moved one step up having taken Seagate ST3160812AS with it. It is pretty strange but Seagate ST3160812AS turned out the only Seagate HDD fighting for the top prize this time, the other three models stayed somewhere in the middle of the race. It is interesting however but the cache buffer was no longer as crucial when working with larger files as it was for smaller files: look at the minimal performance difference between ST3160811AS and ST3160211AS HDDs.
Maxtor’s HDDs have two top places when creating the MP3 file-set, with a small lead over the others. The Seagate ST3160812AS stubbornly holds on to its third place, leaving fourth position to the Hitachi HDS721616PLA380.
Now let’s see if it’s any different when the HDDs have to deal with small-size files.
The Hitachi HDS721616PLA380 is far ahead of the others when creating small files. Its speed is higher by 5MB/s than that of its closest rival Seagate ST3160812AS. The ex-leader Maxtor 6V160E0 finds itself third here.
Let’s now try to read those file-sets:
The Maxtor 6V160E0 is on top again! Without having a high access time or a high sequential read speed, this HDD handles files with ease anyway.
The Hitachi HDS721616PLA380 is second. Files in the Install pattern must be not large enough for the Hitachi to show its speed.
Next goes the reading of largest files from ISO pattern.
That’s right. This diagram almost copies the sequential read diagram from WinBench 99. It’s all about brute force here.
Now the average size of files is reduced:
The Hitachi HDS721616PLA380 has the best speed here too, but the Seagate team has slowed down.
Well, two HDDs from Seagate do better in this test, delivering high speeds and ousting the Maxtor 6G160E0 from third place.
Copying goes next.
Copying the Install file-set into a folder on the same partition is performed faster by Seagate’s ST3160812AS and ST3160811AS HDDs. This diagram shows clearly how the size of the cache affects the write speed. The main point of difference between the two leaders and the last HDD is the amount of cache memory.
As the file size grows up, the gaps between the leaders and outsiders grow wider. The Hitachi HDS721616PLA360 has slowed down much. Firmware of Hitachi HDDs is traditionally good at processing small files, and we’ve got the opposite situation here.
Now let’s make the files smaller.
There’s only one Seagate left on the podium now. The Samsung HD1600JJ, which has never been higher than the middle of the diagram, is second here. The Hitachi HDS721616PLA360 is third.
The Hitachi HDS721616PLA360 has no rivals at copying the smallest files. Its copying speed is much higher than that of its opponents. Seagate’s single-platter HDDs take second and third places.
Now the file-sets will be copied into another partition.
Copying the Install file-sets shows superiority of the Seagate ST3160812AS over its opponents. At a big distance from it and with minimum gaps between each other there are three HDDs: Seagate ST3160811AS, Samsung HD160JJ, and Maxtor 6V160E0.
The leader hasn’t changed, but there are only two pursuers left. Their results are identical while the third pursuer, the Samsung HD160JJ, is slow here and takes seventh place only.
The Samsung HD160JJ can’t make up its mind as to where it belongs. Having a below-average result in the previous diagram, it almost overtakes the leading Seagate ST3160812AS here.
The Seagate ST3160812AS proves its superiority again at copying small files. The Hitachi HDS721616PLA360 is second here.
That’s all about FC-Test and NTFS. If you are interested in the HDDs’ behavior in FAT32, follow this link .
We ran PCMark04 ten times and then averaged the results.
Here’s the diagram:
The Hitachi HDS721616PLA360 takes first place which is no surprise as this HDD has been good throughout our test program. The second- and third-best drives are somewhat unexpected as the WD1600JS and the Samsung HD160JJ have but seldom had top places in the diagrams before, especially the drive from WD which was any good in IOMeter only.
Let’s discuss each of the subtests now.
The System Startup subtest is won by the Hitachi HDS721616PLA360 which enjoys a big lead over the others. The WD1600JS takes second place with optimistic performance, and the Seagate ST3160811AS is third.
The WD1600JS is ahead of all when loading applications and is followed by a Hitachi, but not the Hitachi you might expect. It is the older HDT722516DLAT360 model.
Two HDDs are considerably faster than the others at copying files. These are the Seagate ST3160812AS, followed by the Samsung HD160JJ. There are five more drives behind them that have similar results.
The General Usage subtest confirms Hitachi’s superiority. The company’s drives are both at the top of the diagram. The WD1600JS is the closest pursuer.
We use PCMark05 in the same way as we use PCMark04. Unfortunately, we’ve began using it not so long, and we don’t have data on the two rather old HDDs Samsung SP1614C and Seagate ST3160827AS.
The pair of Hitachi HDDs is on top. The WD1600JS, which has taken a place on the podium in the previous test, is only fourth here. The Maxtor 6V160E0 is third.
Now let’s discuss each of the subtests.
The Hitachi disks have two top places and are followed by the WD1600JS again. Interestingly, the new Startup subtest has a more favorable opinion of the HDT722516DLA380 model which has been at the bottom of the diagram in the same subtest of PCMark04.
The Application Loading subtest doesn’t bring about any surprises. The WD1600JS is first and is followed by the two Hitachis. The Seagate ST 3160211AS is at the opposite end of the diagram – it is the only HDD in this review to have a 2MB cache.
Hitachi’s disks win the General Usage subtest, pushing the WD1600JS to third place again.
It’s hard to find any logic in the results of the Virus Scan subtest, unfortunately.
Maxtor’s disks have two top places in the File Write subtest. The Hitachi HDS721616PLA380 is third and followed by the Seagate team.
That’s the end of our test program, we’ll do some summarizing now.
We’ll write about each manufacturer so that we don’t forget anything.
Hitachi: the HDS721616PLA380 is the better of the two HDDs from this company. This is expectable. It is a single-platter drive with higher platter density. The HDDs had similar results in the synthetic IOMeter test and kept at the middle of the diagrams except for the Web Server pattern when the HDT722516DLA360 was first and the HDS721616PLA380 third. In WinBench 99 the HDS721616PLA380 is the overall leader whereas the HDT722516DLA360 is rather an outsider. In FC-Test these HDDs worked well with small files only. PCMark04 and 05 are not the tests you should judge a HDD by, yet the drives from Hitachi did well in them.
Maxtor: the two HDDs from this company have gone hand in hand through this review, but the 6V160E0 is surely the better of the two. It is in the multi-threaded tests that these HDDs showed their very best – they just met no competition there. In other tests these HDDs had average or good results, except for very poor performance in the IOMeter patterns.
Samsung: Samsung’s disks were close to each other and to the leading group in IOMeter, but then they split apart in the multi-threaded tests. The SP1614C had rather good results at reading, and the HD160JJ was at the bottom of the diagrams, but they changed places at writing. Both showed a good access time in WinBench 99, but didn’t do well in that benchmark afterwards.
These HDDs don’t have good results in FC-Test and in PCMark04/05, but the HD160JJ is somewhat better than the SP1614C.
Seagate: this manufacturer had more HDDs in this review than any other, but three out of its four drives are made on identical platters and differ only in firmware and in the amount of cache memory. And it is the small cache that played a bad trick with the ST3160211AS – it was obviously slower than its mates, let alone the other manufacturer’s HDDs, in most tests.
As opposed to it, the ST3160811AS with the “right” amount of cache memory was always competing for top places. The ST3160812AS had good results, too.
There is one downside, though. Seagate’s HDDs failed completely in the multi-threaded reading test. And the rather old ST3160827AS was the only one of them to show good results in the multi-threaded writing test (due to a stronger server-oriented optimization?)
Western Digital: the single HDD from this firm does well in IOMeter and in both versions of PCMark. In the rest of the tests this HDD was in the middle or bottom part of the diagram.
If we were asked to single out one HDD from those that have been tested here, we would point at the Hitachi HDS721616PLA380. This HDD combines good physical parameters with an ability to put them to good use. It has been among the leaders, or at least not much worse than the leaders, in most of our tests.
Soon we’ll publish a newer edition of the battle of 160GB where the Hitachi drives will be opposed by new models from Samsung, Seagate and Western Digital. That’s when we’ll make our final conclusion.