Over the last several years, the need for greater storage capacity on our hard drives has steadily increased. Not too long ago, a 2 GB hard drive was considered more than adequate to meet our needs, but with the latest operating systems and applications now consuming over 800 MB’s each, an 2 GB drive would quickly become full and is simply inadequate for our modern day needs. Luckily, the hard drive manufactures have done a pretty good job of keeping pace with the increased demand for storage space. Thanks to continued innovations resulting in increasing storage capacity, along with higher data densities, a new 25 GB hard drive can be bought today for the same price a 2 GB drive sold for 3 years ago.

When selecting a hard drive for your system, the two choices today are IDE or SCSI. With lower average seek times and higher rotational speeds than IDE drives, the SCSI drives have always held the edge in performance. This coupled with the other advantages the SCSI interface has over the IDE interface, such as simultaneous access of several drives at once (IDE can only access 1 drive per channel at a time) and the ability to support more devices (IDE can support 4 devices whereas SCSI can support 7 to 15 depending on the interface card), SCSI has always been the choice of high end users. Although SCSI has many benefits over IDE drives, there remains one major drawback. The cost of a SCSI system can easily be double or triple the cost of an equivalently sized IDE system.

So unless you absolutely need support for more than 4 devices, or need access to several devices simultaneously, such as in a server or high end graphics or video editing workstation, for most people the high cost simply cannot be justified. Of course if you have the means, and less than maximum performance is unacceptable to you, then by all means choose SCSI. For most home users, however, with one or two hard drives, and perhaps a DVD player and CD recorder, IDE can offer good performance at a fraction of SCSI’s high cost.

While both IDE and SCSI drives have seen numerous improvements over the last year, with recent the improvements to IDE drives, such as higher data densities, lower seek times, increasing the rotational speed to 7200 RPM and the addition of the Ultra ATA 66 interface standard, IDE drives remain a very competitive option for the small business or home user. Since we are concerned with what the majority of our readers are going to be looking for as far as upgrading, we have chose to stick with IDE drives for this initial round-up. For all of you high-end users out there, don’t fret, as we will have some upcoming SCSI comparisons in the months ahead.

Before we go too far, let’s take a moment to look back at the IDE interface and some of the changes it has undergone over the last several years.

The Evolution of IDE

The IDE or Integrated Drive Electronics interface was developed back in 1988 and was actually an attempt to consolidate all of the different non-SCSI drives at the time to a standard interface. Before that time, many manufacturers loosely followed the designs of CDC interfaces, but they often made proprietary changes that made compatibility difficult. The industry, knowing they needed a standardized interface, formed what is known as the Common Access Method Committee (CAMC), and formulated the specifications for the AT Attachment or ATA standard. In those early days, the ATA interface referred to the controller where IDE referred to the actual drive. This was because there were two other proposed standards at the time, the XT IDE and the MCA IDE standard, both of which never gained industry-wide support. Due to the ATA IDE specification gaining popularity, the other two interfaces quickly died out, and thus, today, the terms ATA and IDE have become loosely interchangeable.

The original AT Attachment (ATA) specification, set forth much of the foundation for today’s Ultra ATA/66 standard. The standard calls for a 4-pin Molex power connector, which supplies +5V dc, +12V dc and ground to the drive, and a 40-pin data/control signal cable. Also, each IDE channel can support 2 devices in a master slave configuration. Although this configuration set the basics for the standard, the design has changed slightly with each new revision, mainly with the actual signal designations on the IDE cable. Note that the original ATA specifications could support 2 devices, or one channel, the later EIDE specification increased support for up to 4 devices, or two channels. Also note that while the Ultra ATA/66 specifications calls for a 80 pin IDE cable, only 40 pins are actually used for data and control signals, the second set of 40 pins are paired with the originals and act as ground wires.

Let’s take a look at some of the major revisions of the ATA standard over the years and look at some of the more important changes that directly relate to performance. The following chart lists the PIO (Programmed I/O Modes), the interface they were used with, and the maximum transfer speeds each mode supported.

Note also the cycle times in the above chart. These times signify the amount of time in which the data has to travel to or from the drive. Since the IDE cable uses 16 wires for data, or is 16 bits wide, PIO Mode 0 can transfer 16 bits of data once every 600 ns. To go into a bit more detail here, a nanosecond (ns) is 1/1,000,000,000 or 0.000000001 of a second, so 600 ns is 0.0000006 of a second. Doing a bit of multiplication we find that 2 Bytes (16 bits) of data is being sent at every 0.0000006 of a second, or 1 byte per 0.0000003 of a second. Taking the inverse of this, or 1 / 0.0000003, we come up with our theoretical maximum transfer rate of 3,333,333.33 Bytes per second or more commonly 3.3 MB/sec (using 1 MB = 1,000,000 Bytes). Using this same formula you can see how the maximum transfer rates of the various modes are determined.

Remember that the cycle times call for the maximum amount of time 2 bytes of data have to be transmitted from the drive to the controller, and that the maximum transfer rate is simply derived from assuming that data is being constantly transmitted. This is often where confusion sets in, as sometimes there are pauses between data transmissions, as it is rare you need 66 MB of data at once, or the drive itself pauses as it seeks the information on an adjoining track. Thus seeing actual transfers of 66 MB/sec are rare, or even impossible on modern IDE hard drives.

Ultra ATA/66

Most of the IDE drives currently on the market are Ultra ATA/66 compliant now that the interface has become the new standard, surpassing the 3-year-old Ultra ATA/33 specification. As all ATA specifications call for backward compatibility, your new Ultra ATA/66 drive will work in older systems. This also means that older PIO and Ultra ATA/33 drives will maintain compatibility with newer Ultra ATA/66 hardware, as well. The requirements for running your Ultra ATA/66 hard drive at Ultra ATA/66 speeds, are an Ultra ATA/66 capable motherboard or IDE controller card and that you use an 80-pin IDE cable instead of the older 40-pin version.

While the pin designations remain the same as with regular IDE cables, at Ultra ATA/66 speeds signal quality issues become a major concern. It is due to this fact that your system needs to determine if you are using the newer 80-pin cable, or older 40-pin cable, before it will enable Ultra ATA/66 mode. While the pin designations are the same, as stated above, one of the lines is broken, where in the 40-pin cable, the connection is unbroken. It is this broken connection that the system will pick up on, to determine if you are using the correct 80-pin cable needed for Ultra ATA/66 operation.

At higher data transfer speeds, cross talk and interference between the data lines impedes the flow of error free communication. To help ensure the integrity of the data signals passing from your drive to the controller, the Ultra ATA/66 standard calls for an addition of a grounding wire to be paired with each one of the signal wires in the interface cable. These 40 grounding wires are only connected at one end and are left un-terminated at the other end to prevent a ground isolation loop from occurring. If you use the standard 40-pin IDE cable with an Ultra ATA/66 drive, your drive will revert to Ultra ATA/33 speeds.

It is due to this “quieter” signal that Ultra ATA/66 is possible. The drive controller must determine whether the data lines have switched to a high or low state before it can read a signal off them. In order to assure a correct reading, the controller must wait awhile after a data burst in order to ensure the signals have settled down to their proper state. This is known as the setup time. With the addition of the 40 extra ground cables, thus filtering out external noise and cross talk, the setup time can in effect be halved. This halving of the setup time is what actually enables Ultra ATA/66 to be so easily implemented without any further significant changes to the Ultra ATA standard.

Now that we have a bit of background, let us introduce you to the drives that are part of this roundup. The drives selected for this round up represent what we consider would be the current choice one would make if upgrading or building a new system today. The only requirements for this roundup were that the drives have 25 GB or more of storage space and that they are compliant with the Ultra ATA/66 standard.

Before we get too far along, let's make a brief recap of what each specification means:

Rotational Speed - The rotational speed, as the name suggests, is the speed at which the platters rotate inside the hard drive enclosure. All else being equal, a faster rotational speed translates into faster read/write performance as the data can be written to, or read from, the platters at a faster rate. This also has the benefit of reducing a drives latency, as the time it takes the requested block of data to arrive under the head is reduced with higher rotational speeds. A disk drives latency is the time it takes the platters to make ½ of a full revolution. Since latency is a direct product of rotational speed, a 5400-RPM drive will have a latency of 5.55 ms and a 7200-RPM drive will have a latency of 4.17 ms. This number is determined by first finding how long it takes the platters to make one complete revolution. First, lets covert the platter speed into rotations per second or 7200/60 (60 seconds per minute) to get 120 rotations per second. Now we determine the time it takes to make 1 revolution by taking the inverse of our rotations per second, in this case 120, so 1 / 120 is 0.008333 seconds for one revolution. Since we need ½ of this time, we divide 0.008333 by 2 to get our latency of 0.0041665 seconds or rounded up to 4.17 ms (1 ms = .001 or 1/1000 of a second).

Cache Memory – As in other parts of your computer, cache memory acts as a buffer to store temporary chucks of data that have been “pre-read” off the drive, or depending on the drive, to hold data that has yet to be written to the drive. The drives use a variety of software algorithms to help predict what data will be accessed next, in determining what to load into the on-board cache memory. If the assumptions are right and you request data that has been pre-loaded into the cache memory, this is also known as a cache hit, your system will save a little time in that it can instantly transfer the data at the maximum burst transfer rate, as it didn’t have to wait for the drive to access it off the platters first. However, this does not help with long sequential transfers of data or in cases where the next requested data was not loaded into the cache (a cache miss); in both cases the drive will have to read data off the disk at normal speeds and no performance increase will be realized. It should be noted that this is where the higher burst transfer rate of an Ultra ATA/66 drive would pay off. If you have a larger cache buffer and the data you requested happened to be all pre-loaded into the cache memory, your data would be transferred at or very near 66 MB/sec. This “data burst” then, is the main advantage to the higher transfer rates. Of course, even with a 2 MB buffer, in long sequential transfers greater than cache size, the cache would be emptied very quickly and you would once again revert to the physical platter read/write transfer rates.

Seek Time – The most common seek time quoted for hard drives is the Average Seek Time or the average amount of time it takes the data heads to perform a random seek of the drive. This is commonly determined by taking ½ of the Full Stroke Time or time it would take the data heads to move from the outermost track to the innermost track or vice versa. Seek times are measured in milliseconds (thousandths of a second) and are often in the range of 7 ms to 11 ms on modern hard drives. Once again, if all other parameters of a drive are equal, the drive with the lowest seek time will offer the better performance. Since seek times relate mainly to the data heads ability to move from one track to another in order to begin reading new data, this specification is important in cases where data is read out of sequence and is spread about on the hard drive.

Capacity – Capacity, which is one of the main things people look for when choosing a hard drive, is the amount of data that can be stored on a hard drive. Capacity is measured in megabytes (MB’s) or millions of bytes and, more recently, gigabytes (GB’s) or billions of bytes. One thing to watch out for when reading the specifications of a drive is whether the number quoted is formatted or unformatted capacity. Sometimes manufactures will quote the unformatted capacity of a drive, which is always higher, but usually it’s the formatted capacity you are looking at when buying a drive.

Another thing that should be pointed out while we are on the subject of capacity, is that when you install your drive into you computer and fire up your operating system, the amount of disk space showing as available is often less than what you were expecting. This is caused by the different methods at which the manufactures and your computer use to compute capacity. Manufactures use a rounding method, which artificially inflates the listed capacity.

An example of this is that 1 KB of data is actually 1024 bytes, 1 MB of data is 1024 KB’s, and 1 GB of data is equal to 1024 MB’s. So 1 gigabyte of data equals 1024 bytes x 1024 KB’s x 1024 MB’s or 1,073,741,824 bytes of data. This is how your computer arrives at its calculation of available storage space. Hard drive manufactures, round these figures down to use an even 1000 in their calculations. This practice, which has become standard, results in 1 gigabyte being equal to 1,000,000,000 bytes or 1000 bytes x 1000 KB’s x 1000 MB’s. So when a manufacturer lists a drive as having 25 GB’s of formatted storage capacity, they are saying you have 25,000,000,000 bytes of storage space and your computer shows you this by listing 23.28 GB’s of space available.

Drive Interface - Even though all of these drives feature the Ultra ATA/66 interface, this interface simply shows the maximum burst transfer rate that the drive can use to communicate with the rest of the system. You should think of the maximum speed of the interface as the theoretical maximum that could be attained and not what the current generation of drives are able to send and receive data at. With the exception of data that is already in the drives cache memory (see above), current hard drives will transfer data at a much slower rate. This rate is often referred to as the maximum sustained transfer rate. The maximum sustained data rate is determined by some of the above factors, rotational speed, on-board data cache and seek times, as well as other factors such as the number of platters and the density at which the data can be recorded. We just wanted to bring this up, as you shouldn’t expect to get double the performance out of an Ultra ATA/66 drive as you would an Ultra ATA/33 unit.

When choosing a hard drive, the three major criteria that you will consider before making your purchase decision are price, desired capacity, and performance. The second two are often dictated by the first, which is price. Since all manufactures are fairly consistent in terms of labeling their hard drives storage capacity, as we outlined above, the main criteria left then is performance. You could simply go by the listed specifications such as those above, but often times those don’t tell the whole story. So now that we have summarized the basic information that you often have access to when shopping for a new hard drive, let’s get on with the comparison and see if these advertised numbers really tell us anything about the drives actual performance.

The Test

The AnandTech storage test bed is currently set up using Windows 98, Windows NT 4.0 Service Pack 6 and Windows 2000. Each operating system is set up using its own 4 GB partition with Windows 98 using the Fat 32 file system, Win NT 4.0 using the NTFS file system and 2 instances of Windows 2000, one using the FAT 32 and the other using the NTFS file system. Since Windows 2000 supports both file systems, we decided to set up two platforms to help compare the performance under both systems. This will help the reader make an informed decision as to which drive performs best under each file system, for users planning to upgrade to Windows 2000. Also, all of the tests were run using the Ultra ATA 66 protocol, using an 80-pin IDE cable. In all instances, the only installed software is the operating system and the benchmark programs.

Some of you may be wondering about our choice of a test bed, most notably the use of a Celeron 500 processor and the Intel 810E motherboard. While this would not make a good choice for the average user looking for maximum performance, as the integrated video graphics leave much to be desired in terms of gaming, it is perfect for use as a test bed. What we are looking for is a solid system, with an integrated Ultra ATA/66 controller, so as to ensure consistent results between tests. The idea here was to eliminate as many variables as possible. Also the 810E supports SDRAM, which was another concern.

It was decided early on to use a motherboard based on the i810E platform because of the fact that the i810E chipset makes use of Intel’s 801AA I/O Controller Hub (ICH) which functions properly as an Ultra ATA/66 compliant controller. The i810E also boasts Intel’s Accelerated Hub Architecture which features a dedicated 266 MB/sec path to the ICH, which helps eliminate any potential bottlenecks that are present with non-AHA chipsets that use the PCI bus as an all-purpose bus used to connect the integrated disk controller with the CPU/memory.  While there are other options such as AMD’s 750 chipset, the performance of a hard drive from one platform to the next should not vary much, so the selection of an i810E motherboard as the test bed should not make a difference.

Using a Celeron 500 as the CPU for our storage test bed was a decision made in order to represent the average of what most users find in their systems as well as because the CPU is sufficient for the needs of our test bed.  Using a faster CPU would not change the outcome of any of our tests.

The complete test bed is as follows:

·         Intel Celeron 500 MHz CPU

·         Intel CA810E motherboard with native Ultra ATA/66 support

·         128 MB Corsair PC100 SDRAM

·        Ultra ATA/66 40-pin 80-conductor HDD cable

·         Microsoft Windows 98

·         Microsoft Windows NT 4.0 SP6

·         Microsoft Windows 2000 (FAT 32)

·         Microsoft Windows 2000 (NTFS)

·         Ziff Davis Winstone 2000 CC

·         Ziff Davis Winstone 99

·         Ziff Davis Winbench 99

·        Adaptec Threadmark 2.

Each hard disk was partitioned and formatted before each suite of tests on their respective operating systems, as to prevent any skewing of the test results. For purposes of consistency, each benchmark was run a total of 5 times, with the final score being the average of those five runs.

Adaptec’s Threadmark 2.0 is a popular benchmark and we used it to help further illustrate the differences between the drives. However, as it is almost 2 years old and is no longer supported by Adaptec; it would not run under Windows 2000, so only Threadmark results for Windows NT and Windows 98 are shown. Ziff Davis’ Winbench 99 was used to show the real-world transfer rates achieved by the individual drives. These two tests are good at demonstrating the sustained data transfer rates of the drives and are useful indicators for the type of applications that access the disk in this type of manner. A good example of this would be digital video and audio editing where data is often accessed sequentially and the raw transfer speed is important. The main consideration for this type of work is going to be a disks rotational speed, as a higher rotational speed will translate into better sequential read/write performance.

Ziff Davis Business Winstone 99 and Content Creation Winstone 2000 were used to show the real world performance differences between the drives. These tests open multiple applications simultaneously and switch between them doing various tasks. These types of tests utilize a lot of random disk accesses and a low seek time in addition to a higher rotational speed and data cache will improve performance here. These tests will more closely represent how the majority of home and business users will access their hard drives and are a useful all around indicator of a hard drives performance.

It should be noted that we could not get all of the commercial tests to run properly under all of the operating systems, and as such you should only compare results across the same platform when comparing the drives. The problems were mainly under Windows 2000 and we don’t know if it was driver problems or if Winstone isn’t 100% Windows 2000 compatible yet. Adaptec’s Threadmark 2.0 refused to even display the drives under Windows 2000, but considering it is a two year old benchmark that is no longer officially supported, it will only be useful for Windows 98 and NT tests.

Here the IBM Deskstar 34 GXP and WD Expert lead the pack, with fellow 7200-RPM drives, the Seagate and Quantum following closely behind. The two 5400-RPM drives fall to the bottom with the Maxtor unit failing to even break the 4 MB/sec mark.

Once again, under the High End Winmark tests, the IBM Deskstar 34 GXP and Western Digital Expert drives turn in identical results and are tied for first place. The Seagate Barracuda comes in a close second, with the 5400-RPM units falling far short of the competition this time.

Things get a bit more interesting here. Using ZD Winbench 99’s Disk Transfer test to measure raw sequential disk transfer rates, the Seagate Barracuda leaves the rest of the field in its dust. You can see that even on the slower inner tracks the Seagate’s performance stays at or above 18.5 MB/sec, while the faster outer tracks reading shows a blazing 28.1 MB/sec. This equates into the actual sustained media transfer rates and you can see where the Seagate Barracuda would be starting to approach the 33 MB/sec maximum transfer rate of the Ultra ATA 33 drive, leaving little room to spare. This illustrates one of the main needs to the move toward the 66 MB/sec speeds of the Ultra ATA 66 standard. While it is useful in burst transfers right now, it leaves plenty of room for growth as the current technology shows that the Ultra ATA 33 spec will very soon become inadequate.

Once again the other 7200-RPM units fill in the middle with the 5400-RPM units pulling up the rear. The IBM 34 GXP and WD Expert continue to post similar results, with the WD slightly pulling out front this time around. It is interesting to note here; that both the 5400-RPM drives outer track transfer rates were very near the Seagate Barracuda’s slowest rate, with the IBM Deskstar 25 GP actually slower. This just further illustrates the importance a higher rotational speed has on a drives performance.

Finally some competition; under ZD Business Winstone 99, the IBM 34 GXP and WD Expert again come in first, but this time around the IBM Deskstar 25GP and Quantum drives fill in the middle. As Business Winstone typically shows the performance of office applications raw transfer rates aren’t as important. Here the benefits of a larger cache and faster seek times come into play, although the results are fairly tight overall.

Things return to normal again under ZD Content Creation 2000. ZD Content Creation simulates typical on-line publishing applications and incorporates a little of both Business and HE tests. Here the 7200-RPM group again separates it self from the slower competition and turn in close score with the Seagate slightly edging out the others.

Although the Adaptec Threadmark 2.0 benchmark is becoming quite old, as it is no longer officially supported by Adaptec, we thought it would be interesting just to toss in the results to give us a further comparison. Here, the WD Expert and IBM Deskstar 34 GXP once again post similar scores, with the Seagate Barracuda taking a nosedive and moving into 4^th place. Now this was really quite surprising, due to the fact that the Seagate Barracuda led the pack under ZD Winbench Disk transfer tests, under Threadmark 2.0 it fell more than a full point from the competition.

Under NT 4, the Quantum unit steps up to the plates and ties with the WD Expert drive for first place in the ZD Business Disk Winmark 99 test. The IBM and Maxtor drives, come in second with scores so close behind the first bunch, that the top four drives are virtually tied. The Seagate barracuda once again surprises us by dropping down the charts, with the IBM Deskstar 25 GP turning in dismal results by itself on the bottom.

Under the High End Disk Winmark test, the Seagate once again rises to the top, followed closely by the Quantum drive this time. The WD Expert and IBM 34 GXP post virtually identical scores again rounding out the middle, with the slower 5400-RPM units once again falling to the bottom of the pile.

Under Windows NT 4, using the ZD Winbench Disk Transfer test, the results are pretty much a carbon copy of the Windows 98 results. The Seagate Barracuda pulls out way ahead of the rest of the group, with the IBM 34 GXP and WD Expert again posting near identical results, as they fill in the second slot. Again, no surprise here, the 5400-RPM units dropped way to the bottom of the chart.

Things are pretty close here, the range isn't significant enough to point out.

Under Windows NT 4, the Adaptec Threadmark results more closely match those of the ZD Winbench Disk Transfer tests. The Seagate Barracuda pulls out ahead of the pack, with the IBM Deskstar 34 GXP and WD Expert once again tying for second place. Predictably the two 5400-RPM drives in the group, the Maxtor DiamondMax and the IBM 25GP, fall to the bottom.

Under Windows 2000 using the FAT 32 file system, the drives perform identically, and finish in the same order, as they did Windows 98 and NT4. The Seagate Barracuda, confirming itself as superior in raw disk transfer speed, leads the group by quite a margin. The IBM Deskstar 34 GXP and WD Expert again turn in identical results for second place.

IBM Deskstar 25GP

The IBM Deskstar 25GP is a 25 GB hard drive that features a 2 MB buffer, 9 ms average seek times, and utilizes the Ultra ATA/66 interface. Unlike the IBM Deskstar 34GXP drive, which is also reviewed here, the 25GP only rotates at 5400-RPM, which lowers it performance significantly and negates any value of using the Ultra ATA/66 interface. As the benchmarks show, only the drives with higher rotational speeds will even come close to needing that amount of bandwidth.

As a result of the lower rotational speed, the IBM Deskstar 25GP was one of the poorer performers in this roundup. With the exception of Business Winstone, the 25GP scored at or near the bottom of all of the tests. This just shows how important higher rotation speeds are when it comes to a drive’s performance. The only test in which the 25GP had even a mediocre showing was in Business Winstone 99. This was likely due to the fact that the 25GP also utilizes a 2 MB cache, like that of the 34 GXP, rather than the 512 KB cache of the other drives. Business Winstone 99 also simulates the common business applications one would use, and as a whole, they typically do not require vast transfers of data. So while this shows that a larger cache does have a slight benefit in certain applications, the main factor to consider when analyzing hard drive performance is still going to be rotational speed.

Again, due to its rotational speed being only 5400 RPM versus the 7200 RPM that is more and more commonly being found on modern IDE hard drives, unless you need the absolute most storage space for your dollar, and performance takes a back seat, we wouldn’t recommend the 25GP especially considering other drives with higher capacities can be found for less money.  If you are after the IBM name, with street prices between the Deskstar 25GP and the higher performing drives only being about $40 in difference, you would be better off going with the one of the other models that offer a extra 2-3 GB’s of storage and a lot more performance for the money.

The IBM Deskstar 25 GP was otherwise an overall solid drive, much like its Deskstar brethren. It ran quietly, with operational and disk access noises barely perceptible, and its temperature never got much above ambient. The 25GP comes with IBM’s standard 3-year warranty and can be found for around $190.00.

IBM Deskstar 34GXP

The IBM Deskstar 34GXP we reviewed is the 27.3 GB version in the 34GXP model line. It features a 2 MB buffer, 9 ms average seek time, Ultra ATA/66 interface, and a rotational speed of 7200 RPM.

The Deskstar 34GXP, as with most of the current drives in the Deskstar family, features IBM's Giant Magneto Resistive (GMR) heads and No-ID sector formatting technology. These two technologies allow an increased amount of data to be stored on each disk surface and reduce the number of platters and heads required for a given capacity. With fewer heads and disks per gigabyte, the drive uses less power and generates less heat and noise than conventional drives.

The performance of the IBM Deskstar 34GXP was right at the top of the charts, neck and neck with the Seagate Barracuda and Western Digital drives and is a very high quality unit. With a outer physical appearance very similar to that of the WD Expert, along with similar technical specifications and almost identical test scores, it is apparent that the IBM Deskstar 34 GXP and WD Expert drives are close cousins, if not the same drive. This comes as no surprise, since WD has been licensing the IBM technology for the last couple of years. Due to the almost identical performance of each drive, I would assume they both come off of the same assembly line and are, in essence, one in the same. Pretty much everything said here about the IBM Deskstar 34 GXP applies equally well for the Western Digital Expert drive.

The 7200-RPM rotational speed really proves beneficial, and with the 2 MB cache memory, the IBM pulls away in the tests that take advantage of a larger buffer. The only test where the IBM 34GXP really fell back was during the raw data transfer tests under Winbench Disk Transfer test. During this test, the Seagate Barracuda really ran away with the show, as it had an average of 4 MB/sec better transfer times across the entire disk surface. However, in the remainder of the tests, the IBM and WD combination dominated the top of the charts.

In testing, the IBM drives were among the quietest of the group; this was especially noticeable during the drive access time tests. While some of the other tested drives sounded like miniature coffee grinders, the IBM drives were barely perceptible in operation. Both drives remained fairly cool to the touch, indicative of a quality product.

Overall, the IBM Deskstar 34GXP line of drives are excellent performers and would be an excellent choice to include in your system. The only real contender to the IBM and WD units is the Seagate Barracuda. The IBM Deskstar 34 GXP 27.3 hard drive comes with a standard 3 year warranty and runs about $240.00.

Maxtor Diamond Max 6800

The Maxtor Diamond Max 6800 hard drive is a 27 GB, 5400 RPM drive with sub 9 ms seek times, and a 2 MB data buffer. It also utilizes the Ultra ATA/66 interface, as do all of the drives tested for this roundup.

The Diamond Max 6800 sports some unique feature such as an onboard dual microprocessor controller, which is supposed to help achieve a theoretical 27.8 MB/sec transfer rate to and from the media. In practice however, the Maxtor’s slower rotation speed hindered its performance, as it didn’t not even approach the claimed data transfer rates.

Some of the other benefits the Diamond Max offers are Maxtor’s “No Quibble Service” policy in which you can call up Maxtor and they will next day air you a replacement drive and you can send back the defective drive at usual ground rates.

In the tests, the Maxtor consistently appeared on the bottom of the charts along with the other 5400 RPM unit, the IBM Deskstar 23GP. While some may think this comparison unfair, the Maxtor did meet our basic specifications, >25 GB capacity and used the Ultra ATA 66 interface. We wanted to demonstrate the importance that the other features of hard drives contribute to the overall performance of a drive, and to further illustrate that simply having an Ultra ATA 66 compatible drive doesn’t necessarily equate into better performance.

Overall, the Maxtor was a fine drive, outperforming the 5400-RPM IBM unit in several tests, but, as we stated with the IBM Deskstar 25GP, unless you are simply looking for the most storage space for the dollar and performance takes second place, then you would be better off looking at a 7200-RPM unit. The days of 5400-RPM drives are surely numbered, as the price differences are so slight that we cannot see choosing one over the higher performance 7200-RPM models.

The Maxtor Diamond Max 6800 comes with a standard 3-year warranty and can be found for around $190.00.

Quantum Fireball Plus KX 27.3 AT

The Quantum Fireball Plus KX is a 27.3 GB hard drive that features a 7200-RPM rotational speed, 8.5 ms average seek times, 512 KB data buffer, and utilizes the Ultra ATA/66 interface.

The Fireball Plus KX also features Quantum’s Data Protection System or DPS. If computer system failures are caused by drive failures, DPS can help a user identify the problem by verifying the drives functionality. This can save unnecessary drive returns as well as protect the users data if the trouble turns out to be non-drive related. The Fireball Plus also features Quantum’s Shock Protection System (SPS), which helps to protect the drive from damage during shipment and installation.

The Quantum drive produced solid results, but failed to stand out among its 7200-RPM competitors. It produced average scores in almost all of the tests and even managed to bubble up to the top on a few, but failed to run away with a decisive victory with any of them. As this drive produced mainly average results in this group, along with the fact that it would have been at the bottom on most of the tests if we had not included the 5400-RPM units, all we can give it is an average rating. If you are looking for an average drive the Quantum will do, but in reality, with the price of the Quantum slightly higher than some of the other high rated drives, who wants to settle for average?

Once again, the Quantum Fireball Plus KX offers average performance in the 7200-RPM, Ultra ATA 66 field but is otherwise an overall solid drive. It ran fairly quiet (but not as quiet as the IBM drives), and its temperature remained at a comfortable level. The Quantum Fireball Plus KX comes with a 3-year warranty and can be found for around $240.00.

Seagate Barracuda ATA

The Seagate Barracuda ATA line of drives is Seagate’s attempt to bring some of the performance of their popular SCSI line into the IDE drive market. Keeping with its SCSI based namesake, the Barracuda ATA offer outstanding performance and an excellent value for you dollar as it outperformed all of the drives in this roundup in raw disk transfer speeds.

The Barracuda ATA ST328040A that we reviewed is a 28 GB hard drive that also features a 7200 RPM rotational speed, 512 KB data buffer, 8 ms average seek times, and utilizes the Ultra ATA/66 interface.

Seagate rates this drive as having a sustained data transfer rate of >15 MB/sec, which was substantiated by our benchmarking. Using the Winbench Data Transfer Test, the Seagate Barracuda ATA delivered 28 MB/sec transfers on the outer tracks and stayed above 18.5 MB/sec on the inner tracks, well above the published claim. In fact the Seagate Barracuda kept a healthy 4 MB/sec lead over the nearest competitors, the IBM Deskstar 34 GXP and Western Digital Expert. Although it ran away with the show in the sustained data transfer tests, the Seagate fell into second place behind the IBM 34 GXP and WD Expert combination.

Since the Seagate really only shined its brightest during the sustained transfer tests, we cannot crown it as absolute victor. Instead we have determined that if you are the type of user that benefits from high-sustained transfer rates, and you use your hard drive for this type of activity often (video and/or sound editing comes to mind), then you would benefit from the Seagate Barracuda and by all means grab it. If you mainly use your hard drive for infrequent access to your data (business apps, general home and gaming use), with occasional sustained transfers, you might be better off going with either the IBM Deskstar 34 GXP or Western Digital Expert drive that we reviewed.

The Seagate Barracuda ATA is a very solid performer and blows away the competition when it comes to sustained data transfers, as it maintained a healthy 4 MB/sec gap from its closest competitor. It is highly recommended if you need a drive that can handle streaming data sessions, such as video editing, and for some reason you want to stick with ATA drives over SCSI. The Barracuda ATA ran a bit warmer than the other drives in operation, we suspect it may be doing a bit more than 7200 RPM, and was a bit noisier than the other drives, especially during the random access tests. The Seagate Barracuda ATA has a 3 year warranty and can be found for around $220.00.

Western Digital Expert 25.7 (WD273BA)

The Western Digital Expert 25.7, WD273BA, is identical to the IBM Deskstar 34 GXP. Besides a few different labels on the drives, you would have a hard time telling them apart if you mixed them up. The similarities don’t end at external looks either, as both drives performed so close in the benchmarks that they had virtually identical results. This really comes as no surprise, being that WD has been licensing IBM’s drive technology for the past few years. Just judging between the similarity in appearance, physical specifications, and performance, we would say WD basically is using the IBM design and just adding their logo.

Since the WD Expert is essentially a clone of the IBM unit, as it shares all the performance benefits that the IBM Deskstar 34 GXP drive offers, we will not go into great detail describing its good points, but rather refer you to the IBM GXP32 summary above. Like the IBM unit, the WD performed at the top of the charts and either one of these drives would make an excellent choice in your system. The decision one would make in choosing one of these two drivers over the other simply comes down to price. Take whichever unit you can get the best deal on.

The Western Digital Expert WD273BA comes with a standard 3-year warranty and can be found for around $230.00, which, as of this writing, is slightly cheaper than the equivalent IBM drive.

Conclusion

Even before beginning to benchmark the different drives, it was expected that the higher rotational speed drives would outperform the slower units. Our testing did not debunk this theory, as the 7200 RPM units all vied for top places in the charts, with the slower 5400 RPM units falling to the bottom of the heap in all but a few of the tests. This isn’t surprising, being that faster rotation speeds equate into faster reads/writes off the platters and offers the added benefit of reduced latency. While the drives incorporate various other features to separate one another, the most bang for your buck, outside of storage capacity, is going to be obtained from a drives rotational speed.

This may lead many to wonder why we even bothered to include the 5400-RPM drives in the roundup. We felt that a great many people are being fed so much hype with the “2x faster sustained data transfer rates” so many manufactures splash all over their boxes when they are promoting the benefits of Ultra ATA 66 over Ultra ATA 33, that we wanted to drive home the fact that simply being an Ultra ATA 66 compatible unit does not automatically make it a faster drive. There are other factors involved as well, most notable among them the rotational speed of a drive.

Our benchmark tests revealed what an impact rotational speed has upon performance, and this will be a more relevant factor in choosing a drive than simply, one uses Ultra ATA 33 and the other Ultra ATA 66. While there are still some differences between the reviewed drives at the same rotational speed, as the results clearly showed, the drives were mainly broken into two distinct categories by their rotational speed. The 5400-RPM units were consistently at or near the bottom, while the 7200-RPM units were always at or near the top.

As for this round up, four of the drives were 7200-RPM units, with the two 5400-RPM units falling far short in terms of performance. As mentioned above, the 7200-RPM units simply send the slower drives to the bottom of the heap, so unless you can get a great deal on one of them and you are simply looking for the most storage space for the dollar, we would recommend going with a 7200-RPM unit. So with that said, and the fact that this review is about performance, we are left to choose among the remaining four contenders. The IBM Deskstar 34GXP and Western Digital Expert 25.7 are virtually clones of one another, so the choice between those two drives simply comes down to price. If you are deciding between these two drives, just grab the one you can get the best deal on.

While the Quantum drive turned in respectable results and held its own in many of the tests, it just didn’t run away with any of them or shine by itself enough to be considered a top candidate. It turned in average results and should be considered an average drive. This, coupled with the fact that it is one of the higher priced drives of the group, leaves it out as a candidate for the top spot. With the IBM Deskstar 34 GXP and WD Expert taking top honors in most of the charts, and with the Seagate Barracuda running away with the disk transfer tests, your overall choice becomes fairly obvious. If you do a lot of video and/or sound editing, or frequently perform other tasks that require a lot of sustained data transfers, the Seagate Barracuda ATA would be the way to go.

The Seagate Barracuda comfortably led in the Disk Transfer category, with a 4 MB/sec lead over the IBM 34 GXP and WD Expert drives. On the other hand, the WD Expert and IBM 34GXP drives led most of the other tests, with the Seagate coming in second in most cases. If you mainly do tasks that require a lot of access to a disk, but not necessarily need continuous sustained data transfers, such as most general applications, gaming included, you would be better off going with either the IBM 34 GXP or WD Expert units. As stated earlier, these two drives are almost identical, and the choice between them comes down to price.

In the future, we plan to have many more storage reviews and we wanted to get the ball rolling quickly, which is why we debuted with only 6 drives. We will also be taking a look at faster offerings from companies such as Maxtor, so they will have a better chance for competing, as we plan to continue to review 7200-RPM, Ultra ATA 66 drives on a regular basis.  We also plan to be taking a look at a few SCSI drives, as well as getting into other storage related devices, such as CDRW and DVD drives.