Reviews Leaderboard Database Reference Search StorageReview Discussion Reliability Survey Search About StorageReview.com Contents

Second Anniversary!
3/13/2000 - SR`s IOMeter Tests
3/14/2000 - The New Database

  March 15, 2000 Author: Eugene Ra  

The results of twenty-one drives tested in StorageReview.com's new testbed serve as the initial building block in an eventual transition to exclusive testing on the new system. Though the benchmark remains unchanged with WinBench 99, the hardware and operating system have undergone significant changes. Here we'll take a brief look at some notable comparisons between old and new WB99 test results.

The assumption is that Windows 2000 Professional's disk accesses and file system resembles NT 4.0 much more than it does Windows 95. As a result, in this article, we're primarily going to focus on the changes between NT4 and Win2k. Remember, certain brands, Maxtor for instance, generally perform better in NT than in Win9x.

First things first... what hasn't changed? Low-Level Measurements such as transfer rates remain constant. Ditto for access times. This is a good thing, of course. One shouldn't expect differences in operating systems or ancillary hardware to affect such scores.

When it comes to high-level scores, however, significant changes abound. The average increase in Business Disk WinMark 99 scores, for example, is 38% when surveying 11 current drives*. In all cases there are increases of at least 25%, ranging all the way up to an impressive 52% gain by the already dominant Atlas 10k II.

Business Disk WinMark 99
Quantum Atlas 10K II
Quantum Atlas 10k II, NT4/Testbed1 - 5423 |
Quantum Atlas 10k II, Win2k/Testbed2 - 8247 |
Maxtor DiamondMax VL20
Maxtor DiamondMax VL20, NT4/Testbed1 - 3807 |
Maxtor DiamondMax VL20, Win2k/Testbed2 - 5577 |
Average SCSI Drive
Average SCSI Drive, NT4/Testbed1 - 4745 |
Average SCSI Drive, Win2k/Testbed2 - 6639 |
Average ATA Drive
Average ATA Drive, NT4/Testbed1 - 4558 |
Average ATA Drive, Win2k/Testbed2 - 6245 |

The story with High-End Disk WinMark scores is a bit different, however. Here the average gain between the old and new testbed is a miniscule 3%, with a few ATA drives actually exhibiting a decrease in measured performance. The Atlas 10k II and, to a lesser extent, the Atlas V are the only drives to exhibit an increase in performance in the High-End WinMark. All other drives posted scores similar to, or slightly lower than, the old system results.

High-End Disk WinMark 99
Quantum Atlas 10K II
Quantum Atlas 10k II, NT4/Testbed1 - 20000 |
Quantum Atlas 10k II, Win2k/Testbed2 - 24633 |
Maxtor DiamondMax Plus 40
Maxtor DiamondMax Plus 40, NT4/Testbed1 - 17167 |
Maxtor DiamondMax Plus 40, Win2k/Testbed2 - 15700 |
Average SCSI Drive
Average SCSI Drive, NT4/Testbed1 - 16741 |
Average SCSI Drive, Win2k/Testbed2 - 18650 |
Average ATA Drive
Average ATA Drive, NT4/Testbed1 - 14304 |
Average ATA Drive, Win2k/Testbed2 - 14024 |

When it comes to individual tests, FrontPage 98 scores jump out, displaying massive gains. Here too Quantum's SCSI drives are the big winners when it comes to gain. Even so, the average gain of our eleven-drive sample is a whopping 111%.

FrontPage 98
Quantum Atlas 10K II
Quantum Atlas 10k II, NT4/Testbed1 - 38133 |
Quantum Atlas 10k II, Win2k/Testbed2 - 99767 |
Seagate Barracuda ATA II
Seagate Barracuda ATA II, NT4/Testbed1 - 44667 |
Seagate Barracuda ATA II, Win2k/Testbed2 - 93750 |
Average SCSI Drive
Average SCSI Drive, NT4/Testbed1 - 37575 |
Average SCSI Drive, Win2k/Testbed2 - 91059 |
Average ATA Drive
Average ATA Drive, NT4/Testbed1 - 43890 |
Average ATA Drive, Win2k/Testbed2 - 84379 |

On the other side of the coin, AVS/Express 3.4 is the villain responsible for keeping the High-End WinMark constant. Here we witness an average drop in scores of 13%. Perhaps this particular application simply runs very poorly in Windows 2000?

AVS/Express 3.4
Quantum Atlas 10K II
Quantum Atlas 10k II, NT4/Testbed1 - 32233 |
Quantum Atlas 10k II, Win2k/Testbed2 - 29567 |
Seagate Barracuda ATA II
Seagate Barracuda ATA II, NT4/Testbed1 - 22567 |
Seagate Barracuda ATA II, Win2k/Testbed2 - 19000 |
Average SCSI Drive
Average SCSI Drive, NT4/Testbed1 - 23317 |
Average SCSI Drive, Win2k/Testbed2 - 21008 |
Average ATA Drive
Average ATA Drive, NT4/Testbed1 - 17795 |
Average ATA Drive, Win2k/Testbed2 - 14928 |

Let's consider the impact that these newer WinBench scores have on currently reigning drives on StorageReview.com's Leaderboard.

7200rpm ATA Drives
Business Disk WinMark 99 High-End Disk WinMark 99
Windows 2000 Professional using NTFS
Maxtor DiamondMax Plus 40 (41.0GB ATA-66) - 7380 |
Seagate Barracuda ATA II (20.4 GB ATA-66) - 7315 |
Western Digital Caviar 7200rpm (20.5 GB ATA-66) - 7037 |
Windows 2000 Professional using NTFS
Western Digital Caviar 7200rpm (20.5 GB ATA-66) - 16800 |
Seagate Barracuda ATA II (20.4 GB ATA-66) - 16300 |
Maxtor DiamondMax Plus 40 (41.0GB ATA-66) - 15700 |

In the 7200rpm ATA stakes, the gap closes considerably between the three current contenders. Maxtor's DiamondMax Plus 40 maintains a tenacious lead over the competition in the Business Disk WinMark. Surprisingly, however, when it comes to the High-End WinMark, the Maxtor drive finishes behind offerings from Seagate and Western Digital. The results are even more startling when one considers that Windows 2000 is an extension of Windows NT, Maxtor's traditional feeding ground. A while back in the StorageReview.com Discussion Forum, readers speculated on whether different drives scale differently as CPU speed changes. Here's more proof affirming such a hypothesis.

5400rpm ATA Drives
Business Disk WinMark 99 High-End Disk WinMark 99
Windows 2000 Professional using NTFS
Western Digital Caviar (30.7 GB ATA-66) - 6450 |
Maxtor DiamondMax 40 (41.0 GB ATA-66) - 6190 |
Quantum Fireball lct10 (30.0 GB ATA-66) - 4900 |
Windows 2000 Professional using NTFS
Western Digital Caviar (30.7 GB ATA-66) - 14967 |
Maxtor DiamondMax 40 (41.0 GB ATA-66) - 13267 |
Quantum Fireball lct10 (30.0 GB ATA-66) - 11533 |

In the case of 5400rpm drives, SR's current leaderboard champion, the Western Digital Caviar 307AA, manages to hold off the others to retain its WinBench 99 performance leadership.

10,000rpm SCSI Drives
Business Disk WinMark 99 High-End Disk WinMark 99
Windows 2000 Professional using NTFS
Quantum Atlas 10k II (36.7 GB Ultra160/m SCSI) - 8247 |
Seagate Cheetah 36LP (36.7 GB Ultra160/m SCSI) - 6170 |
Windows 2000 Professional using NTFS
Quantum Atlas 10k II (36.7 GB Ultra160/m SCSI) - 24633 |
Seagate Cheetah 36LP (36.7 GB Ultra160/m SCSI) - 17167 |

The combination of a faster testbed and Win2k allows Quantum's Atlas 10k II to add even greater distance between itself and the Seagate Cheetah 36LP. The margins here are an astonishing 34% in the Business Disk WinMark and a huge 43% in the High-End.

When it comes to 7200rpm SCSI drives, the Quantum Atlas V currently rests as the only next-generation model reviewed by SR. A review of Seagate's Barracuda 18XL, however, is just around the corner, bringing up the rear of StorageReview.com's 2nd Anniversary Blitz.

From this point onwards, all drive reviews will incorporate the GIF created by WinBench 99 outlining sequential transfer rates across a drive. An interesting contrast can be drawn, for example, between the graphs of the Atlas 10k and Cheetah 36LP. As conveyed by the simple reporting of the Disk/Read Transfer Rate Beginning (40167 vs. 36133), the Atlas 10k II clearly achieves a higher maximum STR. That fast 40 MB/sec speed, however, is maintained by the Quantum drive for a zone that spans only the first 2 gigs of a 36 gig drive. In other words, the Atlas 10k II maintains its top transfer rate along just 5% of its capacity. Contrast this with the Cheetah 36LP, a drive that keeps up its 36 MB/sec STR over the first 9 gigs (25%) of its total capacity. Don't get us wrong- the Atlas 10k II boasts a higher transfer rate than the Cheetah 36LP at any given cylinder. Even so, we believe that users will find the transfer rate graphics useful in discerning differences that aren't otherwise highlighted by the reporting of WinBench 99's STR figures. Below is a pull-down menu with which the reader may examine the STR graphs of all drives tested thus far in the new testbed.

Sequential Transfer Rate Graphs

This wraps up our brief look at the new WinBench99 results. Next up is something that's more revolutionary: IOMeter Performance.

 Some Thoughts on IOMeter Results...

* The 11 Surveyed Drives


HOME | ARTICLES | LEADERBOARD | PERFORMANCE DATABASE | REFERENCE GUIDE
COMMUNITY | RELIABILITY SURVEY | SUPPORT SR! | ABOUT SR |

Copyright © 1998-2005 StorageReview.com, Inc. All rights reserved.
Write: Webmaster