by Alexey Volkov , Nikita Nikolaichev
10/31/2006 | 02:57 PM
Time is unstoppable, but lately it has been going on at such a rate that we seem unable to catch up with it. But it’s even worse for those who are not trying to. Let’s make an attempt then? Let’s first try to recall what has been going on in the hard disk drive market these days. Quite a lot has happened there indeed.
Upon acquiring Read-Rite, a manufacturer of read-write heads, Western Digital has bucked up and is now approaching the industry leader Seagate in terms of market share. It yet cannot beat Seagate in the average price of sold drives because Western Digital doesn’t produce high-profit SCSI drives. However, Western Digital is promoting its very high-performing Raptor drives into the desktop market sector and is quite successfully exploring the notebook drive market with its Scorpio series. Although Western Digital doesn’t use perpendicular recording in its hard disk drives, the capacity of the company’s senior products has reached an impressive 500GB!
Samsung claimed it would become a leader in HDD sales, but they don’t seem to be so confident now. Considering for how long we had to wait for Samsung’s 300GB and 400GB drives to appear, the company has decided to take time and become a leader in the new “perpendicular” world instead.
Hitachi happily digested IBM’s HDD division and released rather successful models of hard drives, but was somewhat wary, just like Samsung, about increasing the areal density per platter. It is only recently that drives on 125GB platters were announced. Until that, Hitachi had only had to do with 80GB and 100GB platters (we mean platters for “desktop” 3.5” drives, of course).
Things of the greatest importance occurred between the two remaining market players, Seagate and Maxtor. In view of the transition to perpendicular recording, Seagate stretched its model range out to the limit. It’s financially burdensome to be putting out three generations of hard disk drives (even four generations if you count in the Barracuda 7200.10 series!) at once especially considering the gigantic production volumes and Maxtor’s dumping on the OEM market. Seagate solved the problem straightforwardly by just buying the rioter. Luckily, Maxtor’s stock had fallen heavily in price after a few unprofitable quarters in a row. With Maxtor acquired, Seagate not only got an opportunity to increase the average price per each sold disk, but also get access to all of Maxtor’s innovations (Maxtor’s SCSI drives are especially fine!) and to Maxtor’s contract for glass platters. Without that contract, Seagate wouldn’t have been able to transfer too many disks to perpendicular recording technology. The problem is that glass platters aren’t being produced in large quantities as yet because there hasn’t been much demand for them. Most companies have been using aluminum substrates.
So, having bought Maxtor, Seagate claimed to have started living in a new, perpendicular world and announced a few products that featured perpendicular recording technology. We’ve already reported to you about the Momentus 5400.3, and today we’re going to check out Barracuda 7200.10.
As for Maxtor, it remained forever in the parallel world. Even the record-breaking areal density of 160GB per platter achieved on DiamondMax 17 drives couldn’t draw the investors back. The dead must be respected, so we’re going to praise Maxtor throughout this article. After all, there are quite a few things it should indeed be praised for.
But the main subject matter of this review is the 750GB model from Seagate’s Barracuda 7200.10 series. Its capacity matters as much as never before.
A long time ago we published a review of the Barracuda 180 drive in which we quoted a funny example from Seagate’s advertising materials: if printed out, the documents stored on the drive would require a pile of paper comparable in height to a skyscraper. We wonder if any skyscrapers five times that height have been built since then? In magnetic data storage, such progress is a reality as you’ll see right now.
As we were writing this review we recalled a small island of Sipadan located near Borneo. This island is the top of a very high cliff that rises up from a depth of 600-700 meters. It is in fact a cliff column about 1500 meters in circumference and rising a little above the sea level. In its top part the cliff is circled with a 50m high live coral reef. The vertical coral walls are the distinguishing feature of the island.
This island, a pearl of Malaysia, offers best places for diving, particularly the Barracuda Point. There are many, very many barracudas there. Indeed, there are thousands of them in there!
Just imagine you are hanging somewhere between the sea surface and the blue abyss. Thousands of fish are spiraling all about you. And then they come… Swift and relentless killers. And you realize there’s only a thin layer of the water costume’s rubber separating you from the hostile environment and there are thousand of wary eyes and jaws around you… Catch your breath now but beware, there’ll be a lot of barracudas in this review, too. We’ll start out with the largest and most dangerous of them!
The table below lists the whole Barracuda 7200.10 series that comprises as many as 25 different models that differ in the interface and cache buffer size as well as in the number of platters and heads. They share one thing in common, though. They all use the perpendicular recording technology (for details on perpendicular recording technology see our article called Seagate Momentus 5400.3 160GB Hard Disk Drive with Perpendicular Recording Technology).
You can note that the right branch of the family tree is shorter by one limb. Seagate decided not to make a 250GB ATA model with a 16MB cache buffer. Well, the user is their judge…
Theoretically, the use of perpendicular recording should have considerably increased the areal density and given Barracuda 7200.10 drives an unheard-of speed of linear reading. Let’s see what speed is provided by the ST3750640AS model that has the highest areal density in the Barracuda 7200.10 series:
It is almost 80MB/s – an astonishing speed indeed!
But the graph has a queerly straight shape. Could Seagate have abandoned adaptive formatting on the transition to perpendicular recording? In theory, such disks should have a reserve in areal density. Let’s check it out. Here’s the zone map for the disk:
So, adaptive formatting is still here. Each head has its own zone distribution. The diagram shows two graphs that reflect the zone distribution for the best and worst heads. In the zero zone, the difference between the surfaces in terms of sectors per track is 110 sectors, which is a big enough difference.
Note also the number of tracks – there are almost 150 thousand of them on the ST3750640AS’s platters!
And what about reading and writing from the cache? Is there any effect from the drive’s support of 3GB/s SATA?
Well, there seem to be none, but we’ve forgot to remove the jumper that limits the speed of the drive to SATA-150. By default, the drives come with this jumper set for better compatibility with older mainboards and SATA controllers.
So, we remove the jumper and run the test once again (we increased the maximum data block size to achieve the highest speed):
And the result is impressive. The max speed of reading and writing from and to the cache is 255.5MB/s and 241.83MB/s, respectively!
Now what about NCQ technology? As usual, we checked it out with IOMeter: the disk was receiving a stream of requests to read random-address sectors. Every 10 minutes the depth of the requests queue was increased.
The drive proves to be slower with enabled NCQ than with disabled NCQ! As we know from our earlier tests (for details see article called Raptor X HDD from Western Digital: Only the "Naked" Truth... ), the Promise controller enables/disables NCQ quite correctly. But we shouldn’t rely on just one controller in such a tricky matter.
Let’s try to use Intel’s controller integrated in the ICH7R chip with AHCI support on and off.
So, the Barracuda 7200.10 looks somewhat better on the Intel controller (with the latest driver). Its performance isn’t brilliant, but at least not a disaster, either. We’ll perform NCQ tests on the Intel controller as on the “more compatible” one.
What else should we check out? Of course, we need to know how NCQ works on write requests. We ran five more patterns on the drive that differed in the share of write requests. Here are the results:
The graph with zero writes (i.e. 100% reads) looks normal, but the rest of them…
It looks like the drive doesn’t manage to put read requests into the queue and defer write requests as effectively as the WD1500ADFD did. So, the problem is on the Barracuda 7200.10 side. We’ll return to this problem later on when we’ll review the Barracuda ES drive.
But we’ve said enough about one drive. It’s time we introduced all the participants of our today’s tests to you.
Here is this fish weighing as much as 750GB. It is codenamed Galaxy 4D where the number 4 stands for the number of platters. Easy to calculate, the data density is an astonishing 187.5GB per platter here!
Descending further along the Barracuda 7200.10 series, we meet the 500GB model codenamed Galaxy 3D. It has three platters, each with a data density of 166GB.
The 500GB model from the Barracuda 7200.9 series employs four rather than three platters because it does not utilize perpendicular recording technology. It is codenamed Tonka 4D and has a data density of 133GB per platter, not impressive by today’s standards.
This NearLine 2 series disk is a very close relative of the previous model. Although the NL.2 series is targeted at professional applications (data storage systems, etc), this relation cannot be concealed even with the “special” NL.2 firmware.
The main feature of this drive is that the manufacturer deliberately slowed down its read/write heads, obviously to make the drive quieter at work. Otherwise, it is a standard representative of the Barracuda 7200.9 series with an 8MB cache buffer.
The school of barracudas ends in the Barracuda 7200.9 model with a SATA interface and a small cache buffer. Considering the abundance of models we have to deal with, you can follow this link to compare their specs easily.
This one is a real Leviathan! The five-platter ten-headed monster from Hitachi has come to meet its opponents. We guess only the Japanese with their belief in the irrational Godzilla can have any hope that the oldie 7K500 will win the battle. Well, we can all get somewhat irrational at times, just like the Japanese.
This ATA-interfaced model differs from the previous one in having a smaller cache buffer (8MB against 16MB). The interface is different, too.
Passed away, but always remaining in our hearts, Maxtor is represented with four models in this review, two from the DiamondMax 11 and two from the MaXLine Pro 500 series.
The drive’s specs are up to the competitors’: four 125GB platters, a 16MB buffer, and a SATA300 interface. Everything a decent hard drive might need to live a successful market life. But the fates decreed different.
As opposed to Seagate, Maxtor didn’t have a habit of slowing down its ATA models by reducing the speed of the heads or the size of the cache. So we can expect Maxtor’s ATA drives to deliver the same performance as their SATA counterparts do.
This series was positioned as disks for entry-level servers and for workstations and thus was directly opposed to Seagate’s NL.2 series. The main feature of the DiamondMax 11 series is the two times higher MTBF, 1 million hours.
It may look strange that the MaXLine Pro series contains a PATA-interfaced model, but we said above that Maxtor didn’t slow down its PATA drives whereas the owner of a storage system shouldn’t care at all which exactly drives from Maxtor he uses.
This is a representative of the new generation of WD drives for desktop computers called Caviar SE16. It has four 125GB platters, a 16MB cache and a jet black coloring.
WD didn’t stay aloof from the common trend to produce professional hard drives based on desktop models. But this is the single company that puts some more or less solid foundation under the division of its products. Thanks to TLER (Time Limited Error Recovery – a limited time to correct a write error) and RAFF (Rotary Acceleration Feed Forward – the drive uses accelerometers to measure vibrations and corrects current disk operations appropriately) and to an increased MTBF (1.2 million hours), WD’s Raid Edition 2 drives can more rightly claim the title of “professional” than any other product in this review.
So, we’ve got quite a lot of drives in this review. Hitachi is represented by two models from the Deskstar 7K500 series. These are also the only five-platter drives in this review. There are four drives from Maxtor here, two models from the DiamondMax 11 and MaXLine Pro 500 series each.
Seagate has the largest presence: three models from the Barracuda 7200.9 series plus the closely related NL35.2 model (it is intended for continuous 24/7 operation in servers and workstations), and two drives from the Barracuda 7200.10 series.
WD provided us with two models, Caviar SE16 and RE2. The former is meant for desktop computers, and the latter, for servers and workstations.
The following table lists the versions of firmware the drives have:
So, the first test we’ll compare the hard drives in is about measuring the average response time of the drive when requested to read or write a sector. The goal of our measurements is to find the read access time and see how aggressive the deferred write algorithms are (roughly speaking, to estimate the number of cache segments allotted to store write requests).
In order to do this, we use IOMeter to bombard the drive for 10 minutes with a stream of requests to read and write 512-byte data blocks with a request queue depth of 1. The drive has to process over 60 thousand requests, so we get a sustained disk response time that doesn’t depend on the amount of cache memory. The results are sorted by the read response time:
The numbers are quite interesting. Hitachi’s drives are in the lead when it comes to reading, and Seagate’s ATA-interfaced drive from the Barracuda 7200.9 series is the last one here. Either to achieve quiet operation of the drive or to separate sharply “fast” SATA drives from “slow” ATA ones, Seagate slows down its ATA models since the Barracuda 7200.7 series.
As for the write response time, Maxtor’s four drives are in the lead and are closely followed by the Hitachi with a 16MB buffer. Seagate’s drives have the worst results, and two of them have downright poor performance with deferred writes. We can explain this for the NL35 drive by Seagate’s desire to increase data storage reliability on server-oriented drives, but this explanation doesn’t work for the ST3500641AS. It must have suffered for the other model (we mean they tested the firmware on it).
Although most of the drives included in this review are not meant to work in servers, these tests had to be performed for two reasons. First, no one can guarantee that users will build RAID arrays only on drives that are specially intended for that. Ordinary desktop drives are commonly employed instead of drives that are “meant for use in disk subsystems of entry-level servers and workstations”. So, while no one can prohibit the user from doing something, the user must be aware that he/she will be responsible for the choice made.
Second, we will see if there are any special optimizations for server-like loads in the server-oriented drives.
Here are the results of the File Server pattern:
The following diagram lists the average of the results of each drive under all kinds of load.
As you see, the two drives from Western Digital enjoy a considerable advantage over their competitors in this pattern. Third place goes to the Hitachi HDS725050KLA360. Seagate’s drives occupy the middle of the diagram, and the two ATA-interfaced drives from Maxtor are the slowest here.
The Hitachi HDS725050KLA360 gains the lead in the Web Server pattern, pushing the WD 5000YS and the WD 5000KS to second and third places. The other drive from Hitachi, the HDS725050KLAT80 model, is fourth. Seagate’s drives, except for the ATA model, again densely occupy the middle of the diagram whereas the Maxtor drives are at the very bottom of it, competing with the ATA drive from Seagate.
We can make two points out of the server load tests. First, the test didn’t reveal any special performance-improving optimizations in the server-oriented drives. We must confess, however, that we tested the drives under comfortable conditions, with a minimum of vibrations and keeping their temperature within 25-30°C. It may be – we have no reason to doubt the manufacturers in this respect – that the specially designed server-oriented drives would have done better under less comfortable conditions.
The second point is that WD’s and Hitachi’s drives were the best in these tests.
This pattern features a large share of write requests which may affect the performance of the drives.
First, we ran the test on the full capacity of the disk.
Like in the previous case, we marked the best results in blue and the worst results in red. Then we calculate the performance ratings for the drives:
We seem to have got some competition now.
Seagate’s drives take up almost all of the top of the diagram. There’s Caviar, a drive from Western Digital, in the school of Barracudas. Well, it’s only the Seagate ST3500630AS that really stands out among the top eight drives. The remaining seven have similar results.
Next we reduce the address space to 32GB and repeat the test once again:
The reduction of the address space affected best the performance of the Hitachi HDS725050KLAT80 which is now considerably better than its opponents. Maxtor’s drives are all noticeably slower than their competitors in this pattern. The other drives have similar results.
We’ll run the threaded tests to check out the drives’ ability to process multiple requests to sequentially located data. This emulates a situation when several large files are being read from the hard disk simultaneously.
To do this, we use IOMeter and time the test so that the amount of processed data was at least two times the size of the drive’s cache buffer. We create up to four simultaneously working threads (each thread is controlled by a worker with a separate disk space range). The disk is accessed in 64KB data blocks and the depth of the outgoing request queue is steadily changed from 1 to 8. The result is the combined read (write) speed of all the running threads.
The number of tables produced is huge, and you can view them here. If you don’t like to analyze them yourself, we can tell you that this test is won by Maxtor’s drives with a colossal advantage over their opponents.
The diagrams below show only the results achieved at a request queue depth of 1 as the closest to the typical load on the disk subsystem of a desktop computer.
When processing one thread, the disks rank up according to their sequential read speeds. No wonder that the first two positions are occupied by Seagate’s 7200.10 drives. The 750GB model is on top since it is the “densest” drive in the series.
What’s curious, we also see Seagate’s drives at the bottom of the diagram (from the 7200.9 and NL35.2 series).
Let’s now see what we have when there are two read threads to be processed:
Well, what’s wrong with the Seagate drives? Can’t they switch to the second gear?
But Maxtor’s drives rise up to the occasion and show their very best. They have almost kept the same speed on switching from one thread to two. That’s what good look-ahead read algorithms can do!
The WD drives have slowed down, but not too much. The Hitachi drives have also become slower, and we can suspect that the smaller speed of the ATA-interfaced model is due to its smaller cache buffer (8MB against the SATA model’s 16MB) – the difference is almost twofold.
Nothing changes with three or four threads. The Maxtors are ahead, the WD drives are doing their utmost to catch up with them, and the Seagate drives are working in random read mode.
Telling you the truth, we can predict the result of this test by ear, without even looking at the screen. If we hear the typical rattle of the read/write heads during this test, then we can be sure the drive won’t show a high speed. But if the drive is silent, its speed is going to be ok.
So, Seagate’s drives have problems here. They do not predict the load and their look-ahead reading is not optimal.
This is the same as the previous test except that we are writing to the disk instead of reading from it.
And we get the same picture as in the previous test: the drives rank up according to their data density. Now, we switch to two threads again…
And there’re no great changes! Well, the Maxtors have pushed the WD drives a little down the diagram, and the two drives from Seagate with 8MB of cache memory have sunk to the bottom (meanwhile, the Hitachi drive with as much of cache shows a much higher speed). But the main thing is that the Seagate drives don’t suffer a catastrophic speed slump here.
The more threads there are, the better positions the Maxtors occupy. The ability to make good use of the cache buffer prevails over the “raw” speed of writing to the platter. By the way, the two Seagate drives with an 8MB cache buffer have improved and outperformed the ATA drive from Hitachi.
So, the threaded tests prove the necessity of having a large cache buffer as well as the ability to make good use of it. Combining both, the Maxtor drives put on a brilliant performance in this test. The awful results of the Seagate drives at processing multiple read threads makes them antiheros of this test.
We use WinBench99 to check out hard disk drives in desktop PC mode. The disk is formatted in NTFS (with the OS tools, using the default cluster size of 4KB) and in FAT32 (with Paragon Partition Manager, using 32KB clusters) for its full capacity. Then we also perform the test over a 32GB partition formatted in NTFS and FAT32 (we create the partition with the standard Disk Manager).
Here are the linear read graphs:
Let’s first compare the drives by the disk access time parameter.
Three drives turn in the best result. These are the two drives from Hitachi, which is not a surprise since all drives of that firm feature a very low access time, and the 750GB monster from Seagate. The latter is a surprise really because the stretched-out platter with a large number of tracks and Seagate’s standard amount of server marks (220) do not improve the random access time.
The only explanation that comes to mind is that the small sample size when measuring the random access has affected the result.
The next diagram shows the speed of linear reading at the beginning and end of the partition:
The Seagate ST3750640AS sports the highest linear read speed both at the fast and slow tracks. The explanation is the same as in the previous case – it has high areal density. The second-best Seagate ST3500630AS has high-density platters, too, but its speed is lower. Moreover, its linear read speed is much lower on the slow tracks than that of the Maxtor 6H500F0. Unfortunately, the drives from Hitachi didn’t perform well here in spite of their having shown a good access time in the earlier tests.
We can now see the results of the drives in the two integral tests, Disk Winmark and High-End Disk Winmark. These tests were performed on a logical 32GB partition.
The WD5000YS boasts the highest High-End Disk Winmark score. Next go the two drives from Hitachi that have identical scores. They are followed by the WD5000KS.
None of the mentioned drives, except for the WD5000YS, could make it to the top three in the Business Disk Winmark test, however. The WD5000YS is third there, and the two top places are taken by Seagate’s ST350063AS and ST3500841AS, respectively.
Now we format the drives in FAT32 and repeat the tests.
You can see that the speeds have got higher, but the positions of the drives in the diagram haven’t changed much. The WD5000KS outperforms the two Hitachi drives in High-End Disk Winmark and takes second place. The top three hasn’t changed at all in Business Disk Winmark.
We’ll use FC-Test in our traditional way. We create two logical 32GB partitions formatted first in NTFS and then in FAT32. We create a set of files on the first partition and then read it, copy it to a folder on the same partition (Copy Near), and then to the second partition (Copy Far).
We start with NTFS. You can see the tables with numbers here. We’ll discuss the results with the help of diagrams.
The drives from Western Digital cope with the task of creating the Install file-set the fastest of all. Third place goes not to the 750GB Seagate as we might have expected from the results of the sequential write test, but to the Seagate ST3500630AS. The ways of the file system are inscrutable to men indeed…
The WD5000KS moves up to the top when creating large ISO-like files whereas the ex-leader WD5000YS sinks to fourth position. The Seagate ST3500630AS took the vacant second place, leaving third position to the Maxtor 6H500F0.
With minor changes, the diagram of creating MP3-like files is almost the same as the diagram for the Install file-set. The drives occupy the same places in the diagram and also have nearly the same speeds.
The two leaders are the same again, but third place is now occupied by the Hitachi HDS725050KLA360. Note also that there is a considerable difference between the top three drives and their competitors. We haven’t seen that in the previous tests. This must be due to the much smaller average file size in this file-set.
Creating the smallest file-set produces a diagram that differs somewhat from what we’ve seen earlier. The WD5000YS habitually wins here, but then there are changes. Hitachi’s drives take second and third positions. The Seagate ST3500630AS finds itself in fourth place. The second drive from Western Digital, WD5000KS, is only fifth.
Now the drives will be reading the same file-sets.
The two drives from Western Digital are again the fastest when working with the Install pattern. Seagate’s ST3750640AS and ST3500630AS are in third and fourth positions, which is somewhat alarming considering that they boasted the highest sequential read speed in the previous benchmarks. Well, reading small files is not exactly like sequential reading. Let’s see how the drives read large files of the ISO pattern.
Here, everything is as it should be. High data density does the trick.
The two new Barracuda 7200.10 enjoy a considerable lead over their opponents. What’s funny, the remaining drives from Seagate are at the bottom of the diagram.
Reading MP3-like files proves that high data density does not necessarily guarantee a success at reading all types of files. Seagate’s drives, except for the 750GB model, are at the bottom of the diagram. And the ST3750640AS could only surpass the two drives from Hitachi. The pair of drives from WD are the leaders and are followed by the Maxtor quartet.
Now let’s see what we have at reading small files.
The WD drives take the leading position again, the WD5000YS model having a rather big advantage. The Maxtor 7H500R0 has good results, too. Its position is always in the top part of the diagram. Seagate’s ST3750640AS and ST3500630AS have restored their reputation to some extent.
Now we are going to copy the same file-sets into a subfolder.
The top three drives enjoy a big lead over the rest when copying the Install file-set. And there are new names there, too. The previously dominating WD5000YS is now only third while the two top positions are occupied by the best models from Seagate, the ST3750640AS and the ST3500630AS.
In the test of copying large ISO-like files, Seagate again claims the top two places for itself. Maxtor’s drives take up the places from third through sixth. The drives from Western Digital have stumbled here and produced obviously poor results.
The files are smaller, and the drives from Western Digital can copy them quickly again. They are again contending with Seagate’s ST3750640AS and ST3500630AS for the top part of the diagram.
The files getting even smaller, the Hitachi HDS7225050KLA360 becomes engaged into the struggle of Seagate’s and WD’s drives. It first removes the WD5000KS from fourth place and then even becomes the leader.
Now we’ll be copying the same file-sets to another logical disk and see if the results are any different.
Copying the Install file-set into another partition is almost like copying the same files into a subfolder. The Seagate ST3750640AS has a solid reserve of speed over the other drives.
It looks like a trend that Western Digital’s drives have low speeds when copying large files. For the Maxtors, it is the best conditions as they are closely approaching the leaders, Seagate’s ST3750640AS and ST3500630AS.
The diagrams that show copying to another partition are very similar to the diagram of copying into a subfolder, so we won’t discuss them further.
The results of the test in FAT32 differ but little from the NTFS results we’ve just discussed. So, they don’t require our comments. You can view the tables and diagrams here.
We discussed the methodology of benchmarking hard disk drives with PCMark04 in the article called PCMark04: Benchmark for Hard Disk Drives?. We ran the benchmark ten times and averaged the results for each of the subtests.
The following diagram is built according to the HDD Score parameter and sorted in descending order.
The Hitachi HDS725050KLA360 drive has the highest performance score, but the WD drives aren’t far worse, taking second and fourth places. The Hitachi HDS725050KLAT80 finds itself in third place.
Now we can have a look at the results of each of the subtests.
This subtest measures the time it takes to boot up the operating system. To our surprise, it is the ATA drive from Hitachi that turns in the best result here. Until now this drive has only been good in the Workstation pattern. The WD5000YS is second. Next goes the drive from Seagate’s new series, and then goes a Maxtor. So, we’ve got one drive from each manufacturer in the top of the diagram. Surely, the manufacturers all pay attention to this operation mode (booting the OS up) when optimizing the performance of their products, but why only one drive from each brand?
The Hitachi drives are the best at loading applications. These drives know how to handle small files. The WD drives aren’t much slower, though. It’s rather funny to see such paired performances.
The new-generation 500GB drive from Seagate wins the subtest of copying files. The other drive from the 7200.10 series is close, too, residing in third position of the diagram. To our surprise the SATA drive from Hitachi has wedged in between them. We are surprised because that drive didn’t show a high copying speed in FC-Test. On the other hand, PCMark’s copying subtest doesn’t copy files. It just plays back on one hard drive a record of another hard drive’s activity.
The last test shows the average performance at general HDD usage. Hitachi’s SATA drive is an obvious leader here. Note also that we’ve got all the drives from Hitachi and WD at the top of the diagram again. The bottom of the diagram is occupied by the Seagate crowd.
As time is passing by, applications and benchmarks grow old and inadequate. It’s time to try PCMark05!
The drives from Hitachi take the two top positions with a considerable lead. There’s only one WD drive following them now, it is the WD5000YS. The other drive from WD is only in eighth position. But the Seagate ST3750640AS is now very close to the WD5000YS.
Now let’s analyze the results of the drives in each of the subtests.
It seems the operating system Futuremark specialists used to record the trace didn’t change much. The two drives from Hitachi win here with a great lead over the others. They are followed by the two drives from WD while the bottom of the diagram is a place of struggle between Maxtor and Seagate.
The SATA drive from Hitachi copes with loading applications the best of all. After it, the two drives from WD and the second Hitachi are contending. The Maxtors are all at the bottom of the diagram, just like they were in PCMark04.
The General HDD Usage subtest is won by the Hitachi HTS725050KLA360, too. And again the drives from Hitachi and WD are at the top of the diagram. The difference from the previous subtest is that all the drives from Seagate are now at the bottom.
New in PCMark, the test of scanning files for viruses produces results that are hard to explain. Can the scanning speed be higher than the sequential read speed? We guess we won’t make any judgments basing on this subtest.
The test that measures the speed of writing a file (or files?) agrees with the earlier obtained results. The drives rank up according to their areal density.
So, that’s the end of our test program. We’ve got to do some summarizing now.
In order not to miss anything, we’ll sum everything up separately for each manufacturer.
Hitachi: the HDS725050KLA360 drive showed good results in IOMeter’s File Server and Web Server patterns, but moderate in the Workstation pattern. The HDS725050KLAT80, on the contrary, didn’t show anything exceptional throughout IOMeter, but took first place in the last Workstation32 pattern.
These drives showed their best in FC-Test at copying small files. Under other types of load, the drives didn’t perform well.
It was in PCMark04 and PCMark05 only that these drives took leading positions.
Maxtor: the drives from this company passed the test program in a dense group and with average results, except for the threaded tests where they had no rivals. Anyway, we want to say goodbye and thanks to the company for its long and fruitful work. There’ll be no more HDDs under this brand.
Seagate: the drives from this company come from two different families: Barracuda 7200.9 and Barracuda 7200.10. And they perform differently, too. The models from the older family have obviously low results whereas the new family with increased data density stands out among the rest. Excellent results were achieved in IOMeter’s Workstation and Workstation32 patterns. They had a good access time in Winbench99 and the highest linear read speed, too. FC-Test showed their superiority at copying medium-sized and, especially, large files.
Western Digital: the two drives from this company were the stars of this review. They have always been in the top of each diagram throughout the entire test program. Their only weak spot is copying large files, but that’s not a great shortcoming considering their victories.
Our tests have also shown that the performance of modern hard drives is affected by the cache buffer algorithms more than by the physical design. This is clearly illustrated by the results of Hitachi’s and Maxtor’s drivers. Neither of them has mega-dense platters, yet they have performed superbly in a number of tests.
Having said our goodbye to Maxtor, we stumble upon one idea… What if the superb physics of Seagate’s drives is combined with the excellent brains of Maxtor’s? The next year may prove to be very exciting in the HDD area!
The barracudas are gone, and we can catch our breath.
But they’ll be back. And they’ll be larger and more blood-thirsty then!