by eugene

Setting a New Standard - A Revised Look at Serial ATA Performance


When we established our third-generation testbed nearly two years ago, serial ATA remained a technology resting on the horizon. Testbed3 incorporated an Intel 850MV motherboard- while it supported ATA-100 natively, we added a Promise Ultra133 controller to test ATA-133 and/or drives that were larger than 137 gigabytes in size.

Seagate's Barracuda ATA V hit the scene approximately one year later. After conducting some preliminary tests with a variety of SATA controllers, we standardized on the Promise SATA150 TX4 host adapter in no small part to maintain consistency with the Ultra133. At the time, however, the latest available BIOS and driver builds were still in a "pre-release" stage.

More SATA units trickled into our lab. To maintain consistency with our first SATA review (the 'Cuda V's), we stuck to the same BIOS and driver. Recently, however, Western Digital sent us an SATA version of their massive 250 GB Caviar, the WD2500JD. While the high-level performance delivered by the JD combined with the Promise controller was still within expectations, the pair exhibited an anomalous transfer rate graph, one that started with low rates that peaked halfway through the drive and then decayed again as expected. Western Digital, in its own internal testing, could not duplicate the problem. They were not, however, using pre-release drivers like we were.

It became clear that, to eliminate the problem, we would have to update our controller's software to the latest versions. Accordingly, we've flashed the TX4's BIOS to version 1.00.0.33 and the Windows XP driver to version 1.0.0.27. Paired with this updated software, the WD2500JD exhibits a transfer rate graph identical to its PATA WD2500JB brother.

To compare the JD (as well as all upcoming drives) fairly with those we've already evaulated, however, a retest of all previously reviewed SATA drives is required. The following figures convey the change in performance exhibited by the 200 GB Maxtor DiamondMax Plus 9, the Seagate Barracuda ATA V, and the Western Digital Raptor WD360GD.

Note that the Barracuda 7200.7 is conspicuously missing. Long-time SR sponsor HyperMicro was gracious enough to supply a loaner unit for our previous review. We're in the process of obtaining another sample and will published revised results when we receive one.


Low-Level Results

For diagnostic purposes only, StorageReview measures the following low-level parameters:

Average Read Access Time- An average of 25,000 random accesses of a single sector each conducted through IPEAK SPT's AnalyzeDisk suite. The high sample size permits a much more accurate reading than most typical benchmarks deliver and provides an excellent figure with which one may contrast the claimed access time (claimed seek time + the drive spindle speed's average rotational latency) provided by manufacturers.

WB99 Disk/Read Transfer Rate - Begin- The sequential transfer rate attained by the outermost zones in the hard disk. The figure typically represents the highest sustained transfer rate a drive delivers.

WB99 Disk/Read Transfer Rate - End- The sequential transfer rate attained by the innermost zones in the hard disk. The figure typically represents the lowest sustained transfer rate a drive delivers.

For more information, please click here.

Note: Scores on top are better.
Service Time Graphs (in milliseconds)
Average Read Service Time
Western Digital Raptor WD360GD (36 GB SATA) - 8.6|
|
Western Digital Raptor (36 GB SATA) (Updated Bios/Driver) - 8.6|
|
Maxtor DiamondMax Plus 9 (200 GB SATA) (Updated Bios/Driver) - 13.5|
|
Maxtor DiamondMax Plus 9 (200 GB SATA) - 13.6|
|
Seagate Barracuda ATA V (120 GB SATA) (Updated Bios/Driver) - 13.8|
|
Seagate Barracuda ATA V (120 GB SATA) - 14.2|
|
Note: Scores on top are better.
Transfer Rate Graphs (in megabytes per second)
Transfer Rate - Begin
Western Digital Raptor WD360GD (36 GB SATA) - 63.1|
|
Western Digital Raptor (36 GB SATA) (Updated Bios/Driver) - 63.0|
|
Maxtor DiamondMax Plus 9 (200 GB SATA) - 53.7|
|
Maxtor DiamondMax Plus 9 (200 GB SATA) (Updated Bios/Driver) - 53.6|
|
Seagate Barracuda ATA V (120 GB SATA) - 43.8|
|
Seagate Barracuda ATA V (120 GB SATA) (Updated Bios/Driver) - 43.6|
|
Transfer Rate - End
Western Digital Raptor WD360GD (36 GB SATA) - 45.0|
|
Western Digital Raptor (36 GB SATA) (Updated Bios/Driver) - 45.0|
|
Maxtor DiamondMax Plus 9 (200 GB SATA) - 31.6|
|
Maxtor DiamondMax Plus 9 (200 GB SATA) (Updated Bios/Driver) - 31.6|
|
Seagate Barracuda ATA V (120 GB SATA) - 24.7|
|
Seagate Barracuda ATA V (120 GB SATA) (Updated Bios/Driver) - 24.7|
|



Single-User Performance

StorageReview uses the following tests to assess non-server use:

StorageReview.com Office DriveMark 2002- A capture of 30 minutes of actual computer productivity use that exactingly recreates a typical office-style multitasking environment. The applications include: Outlook XP, Word XP, Excel XP, PowerPoint XP, Calypso (a freeware e-mail client), SecureCRT v3.3 (a telnet/SSH client), CuteFTP Pro v1.0 (an FTP/SSH client), ICQ 2000b), Palm Hotsync 4.0, Gravity 2.3 (a Usenet/newsgroups client), PaintShop Pro v7.0, Media Player v8 for the occasional MP3, and Internet Explorer 6.0.

StorageReview.com High-End DriveMark 2002- A capture of VeriTest's Content Creation Winstone 2001 suite. Applications include Adobe Photoshop v5.5, Adobe Premiere v5.1, Macromedia Director v8.0, Macromedia Dreamweaver v3.0, Netscape Navigator v4.73, and Sonic Foundry Sound Forge v4.5. Unlike typical productivity applications, high-end audio- and video- editing programs are run in a more serial and less multitasked manner. The High-End DriveMark includes significantly more sequential transfers and write (as opposed to read) operations.

StorageReview.com Bootup DriveMark 2002- A capture of the rather unusual Windows XP bootup process. Windows XP's boot procedure involves significantly different access patterns and queue depths than those found in other disk accesses. This test recreates Windows XP's bootup from the initial bootstrap load all the way to initialization and loading of the following memory-resident utilities: Dimension4 (a time synchronizer), Norton Antivirus 2002 AutoProtect, Palm Hotsync v4.0, and ICQ 2000b.

StorageReview.com Gaming DriveMark 2002- A weighted average of the disk accesses featured in five popular PC games: Lionhead's Black & White v1.1, Valve's Half-Life: Counterstrike v1.3, Blizzard's Diablo 2: Lord of Destruction v1.09b, Maxis's The Sims: House Party v1.0, and Epic's Unreal Tournament v4.36. Games, of course, are not multitasked- all five titles were run in a serial fashion featuring approximately half an hour of play time per game.

For more information, please click here.

Note: Scores on top are better.
Desktop Performance Graphs (in I/Os per second)
SR Office DriveMark 2002
Western Digital Raptor WD360GD (36 GB SATA) - 500|
|
Western Digital Raptor (36 GB SATA) (Updated Bios/Driver) - 500|
|
Maxtor DiamondMax Plus 9 (200 GB SATA) (Updated Bios/Driver) - 385|
|
Maxtor DiamondMax Plus 9 (200 GB SATA) - 377|
|
Seagate Barracuda ATA V (120 GB SATA) - 366|
|
Seagate Barracuda ATA V (120 GB SATA) (Updated Bios/Driver) - 364|
|
SR High-End DriveMark 2002
Western Digital Raptor WD360GD (36 GB SATA) - 524|
|
Western Digital Raptor (36 GB SATA) (Updated Bios/Driver) - 495|
|
Maxtor DiamondMax Plus 9 (200 GB SATA) - 395|
|
Maxtor DiamondMax Plus 9 (200 GB SATA) (Updated Bios/Driver) - 379|
|
Seagate Barracuda ATA V (120 GB SATA) - 355|
|
Seagate Barracuda ATA V (120 GB SATA) (Updated Bios/Driver) - 332|
|
SR Bootup DriveMark 2002
Western Digital Raptor WD360GD (36 GB SATA) - 526|
|
Western Digital Raptor (36 GB SATA) (Updated Bios/Driver) - 465|
|
Maxtor DiamondMax Plus 9 (200 GB SATA) (Updated Bios/Driver) - 358|
|
Seagate Barracuda ATA V (120 GB SATA) - 338|
|
Maxtor DiamondMax Plus 9 (200 GB SATA) - 332|
|
Seagate Barracuda ATA V (120 GB SATA) (Updated Bios/Driver) - 305|
|
SR Gaming DriveMark 2002
Western Digital Raptor WD360GD (36 GB SATA) - 625|
|
Western Digital Raptor (36 GB SATA) (Updated Bios/Driver) - 614|
|
Maxtor DiamondMax Plus 9 (200 GB SATA) - 535|
|
Maxtor DiamondMax Plus 9 (200 GB SATA) (Updated Bios/Driver) - 512|
|
Seagate Barracuda ATA V (120 GB SATA) - 480|
|
Seagate Barracuda ATA V (120 GB SATA) (Updated Bios/Driver) - 460|
|

In the SR Office DriveMark 2002, though the Raptor and Barracuda hold steady at their previous scores, the DiamondMax Plus 9 delivers a small but, given the statistical precision of the DriveMark playbacks, significant gain to 388 I/Os per second from its previous 377 IO/sec score. Though a 2% gain is not exactly much to write home again, it demonstrates that hard drives, controllers, and their various software versions all interact to deliver synergetic differences.

The High-End DriveMark 2002 reveals a universal slide among all three players. Paired with Promise's beta firmware and drivers, the Raptor, for example, managed to rival and even exceed current-generation 15,000 RPM SCSI drives. With the latest releases, however, WD's drive slips to 495 I/Os per second- while nothing to sneeze at, the revised score permits the 15k units a clean segregation from SATA drives (the Seagate Cheetah 15k.3 holds the "low" 15k score of 495 I/Os per second; though not displayed in these graphs, readers may access custom comparison data through the Performance Database). The DiamondMax and Barracuda exhibit a similar 4-6% slide in High-End scores.

The Raptor demonstrates a precipitous slide when it comes to the SR Bootup DriveMark 2002, a 12% fall from 526 I/Os per second all the way down to 465 IOs/sec. Interestingly, though Seagate's drive exhibits a similar loss, the Maxtor manages to turn things around and gains 8% on its bootup score with the new software.

Finally, when it comes to the SR Gaming DriveMark 2002, a weighted average of five different PC games from varying genres, all three drives take a slight step back, ranging from just 2% in the case of the Raptor to about 4% for Maxtor's and Seagate's units.


Multi-User Performance

StorageReview uses the following tests to assess server performance:

StorageReview.com File Server DriveMark 2002- A mix of synthetically-created reads and writes through IOMeter that attempts to model the heavily random access that a dedicated file server experiences. Individual tests are run under loads with 1 I/O, 4 I/Os, 16 I/Os, and 64 I/Os outstanding. The Server DriveMark is a convenient at-a-glance figure derived from the weighted average of results obtained from the four different loads.

StorageReview.com Web Server DriveMark 2002- A mix of synthetically-created reads through IOMeter that attempts to model the heavily random access that a dedicated web server experiences. Individual tests are run under loads with 1 I/O, 4 I/Os, 16 I/Os, and 64 I/Os outstanding. The Server DriveMark is a convenient at-a-glance figure derived from the weighted average of results obtained from the four different loads.

For more information click here.

Note: Scores on top are better.
Server Performance Graphs (in I/Os per second)
SR File Server DriveMark 2002
Western Digital Raptor (36 GB SATA) (Updated Bios/Driver) - 196|
|
Western Digital Raptor WD360GD (36 GB SATA) - 193|
|
Seagate Barracuda ATA V (120 GB SATA) (Updated Bios/Driver) - 119|
|
Seagate Barracuda ATA V (120 GB SATA) - 114|
|
Maxtor DiamondMax Plus 9 (200 GB SATA) (Updated Bios/Driver) - 112|
|
Maxtor DiamondMax Plus 9 (200 GB SATA) - 96|
|
SR Web Server DriveMark 2002
Western Digital Raptor (36 GB SATA) (Updated Bios/Driver) - 194|
|
Western Digital Raptor WD360GD (36 GB SATA) - 183|
|
Seagate Barracuda ATA V (120 GB SATA) (Updated Bios/Driver) - 126|
|
Maxtor DiamondMax Plus 9 (200 GB SATA) (Updated Bios/Driver) - 120|
|
Seagate Barracuda ATA V (120 GB SATA) - 119|
|
Maxtor DiamondMax Plus 9 (200 GB SATA) - 116|
|

Under load conditions that vary from a single queued operation all the way up to 64 outstanding I/Os, all three drives from all three manufacturers experience across the board gains in both the File Server (a mix of reads and writes weighted mostly towards the former) and Web Server (a mix of pure reads) patterns. The File Server and Web Server DriveMarks, weighted averages of individual tests, reflect these gains. The most notable arises in the DiamondMax's File Server performance- a respectable 17% gain.


CPU Utilization

The CPU utilization of today's modern storage subsystems is determined almost entirely by the controller/driver combo. Whether paired with Maxtor's, Seagate's, or WD's drive, the old BIOS/driver combo delivered CPU utilization hovering around 17% at 10,000 requests per second on a 2 GHz P4. The revised controller system's score rises to about 23-24%. Higher in an absolute sense, yes, but all in all, CPU utilization of a single drive or even small RAID arrays is not a factor in today's PCs- one will never hit the level of I/Os per second necessary for the storage system directly impact CPU performance in a significant manner.


Conclusion

Where do we go from here? The "standard" performance figures for these three drives will be formally updated in the performance database shortly.

With this newly revised baseline with which to compare upcoming reviews, Testbed3 stands ready to evaluate a new crop of exciting units just around the corner. Western Digital has thrown its hat into the 7200 RPM SATA arena with the WD2500JD. Maxtor has incorporated the SATA interface on its high-capacity DiamondMax Plus 9 series. Hitachi has announced the Deskstar 7K250. We know many readers can't wait to see how these new player stack up. Neither can we!

  Review Discussion