iXBT Labs - Computer Hardware in Detail

Platform

Video

Multimedia

Mobile

Other

Seagate U6 HDD review

August 14, 2001



Seagate's share on the hard disc market is extending more and more, partially at the expense of reorganization of Quantum and Maxtor. The company started with a revolutionary disc Cheetah X15 (RPM 15000) and then turned to Barracuda 180 and Barracuda 36ES. But this process involves not only high-efficient SCSI discs but also IDE devices. As you know, Barracuda ATA takes one of the leading places among 7200 RPM discs, and Seagate U5 boasts of a high speed and a low price.

But U5 series is not the final one: recently it was replaced with the Seagate U6 line. At the company's site this series is often called simply Seagate U. The basic difference of this line from the previous one is twice extended record density. Now one side of the disc can contain 20GB! Besides, in the new series Seagate extended the buffer volume on the discs up to 2 MBytes. In appearance, the new disc of Seagate doesn't differ from the U5 one (the bottom side is covered with a SeaShield, but we have taken it off for a photo):

Seagate U6 front (31k)

Seagate U6 back (31k)

The new Seagate's series is comparable to Fujitsu MPG3AT, Maxtor DiamondMax VL40 and Western Digital WDx00AB. Besides, we have taken the Seagate U5 disc for the tests.

Technical characteristics of the competitors:

Series Seagate U6 Seagate U5 Fujitsu MPG3AT Maxtor DiamondMax VL40 Western Digital WDx00AB
Model number
ST380020A
ST360020A
ST340810A
ST340823A
ST330621A
ST320413A
ST310211A
MPG3409AT
MPG3204AT
MPG3102AT
34098H4 33073H3 32049H2 31024H1
WD800AB
WD600AB
WD400AB
WD300AB
Formatted Capacity (GB)
80/60/40
40/30/20/10
40/20/10
40/30/20/10
80/60/40/30
Internal Transfer Rate (max) (Mbits/sec)
436
376
398
374
424
Average Seek Time, Read (ms)
8.9
8.9
9.5
9.5
9.5
Track-to-Track Seek, Read (ms)
1.2
1.2
1.0
1.0
2.0
Average Latency (ms)
5.55
5.6
5.56
5.55
5.0
Buffer Size
2 MBytes
512 KBytes
512 KBytes/2 MBytes
2 MBytes
2 MBytes
Spindle Speed (RPM)
5400
5400
5400
5400
5400
Discs/Heads
2/4
2/3
1/2
2/4
2/3
1/2
1/1
2/4
1/2
1/1
2/4
2/3
1/2
1/1
3/6
2/4
2/3
1/2
Track Density (TPI)
58,000
31,100
33,000
Recording Density (BPI, max)
562,436
477,800
478,415

Notes:

1. Seagate ST315311A for 15 GBytes and Fujitsu MPG3307AT for 30 GBytes are absent in the table.

2. Different companies use different techniques of measuring Average Seek Time. No other explanations why equally named parameters do not correspond to each other are known.

The record density is higher because of the increased number of cylinders on a disc (and not at the expense of sectors on a track). This way was taken by Seagate as relatively cheap for extending a disc volume - it's the first time when a 80 GBytes disc uses only two platters. The access time have probably grown (that is why IOMeter results may be worse). But the increased disc volume should positively tell upon the Winbench test. But let's wait and see...

The testing technique and test platform completely coincide with the earlier used ones.

The competitors in the test are Seagate U5 ST340823A, Fujitsu MPG3204AT, Maxtor DiamondMax VL40 34098H4, Western Digital Caviar WD300AB.

Let me start with the ZDLabs Winbench.

The graph of linear reading from a disc surface (the horizontal axis shows percentage of a disc volume):

Comparison with the contestants:

As compared with the U5 model the linear reading speed has decreased a little bit at the beginning, and the ST380020A joins to a wide group of storage devices with the highest reading speed about 30 MBytes/sec. And as for the reading speed at the end of the disc, the Seagate U6 takes the last place.

The access time has grown a bit as compared with the U5 series. It should affect the IOMeter test which is very sensitive to this parameter.

Comparison with the integer parameter of the Winbench looks much more interesting:

In all tests, apart from the Business with FAT32, the new Seagate takes a firm third place. As far as the strong fall in the first group is concerned, the disc seems to have suffered from its large volume. The FAT32, on the one hand, allows using hard discs up to 128 GBytes. But on the other hand, FAT itself grows as the disc volume gets larger. It means that the time of operating with it also becomes more. Of course, we may increase a cluster size, but in this case time of operation with small files becomes higher, and losses of the disc volume grow. Let's look at the size of one FAT table for a FAT32 partition of 40 GBytes, and at an average size of losses if it contains 20K files:

cluster size (KB)
2
4
8
16
32
FAT size (MB)
80
40
20
10
5
losses (MB)
20
40
80
160
320

We may also compare parameters of the Business Disk WinMark for a 30 GBytes partition being formatted for the FAT32 with different cluster sizes:

As you can see, the results depend much on a cluster size. Unfortunately, the developers of this test do not give its detailed description. It seems that they use a very simple template, for example, operation with 8 KBytes blocks, 20% random access/80% serial, read only. That is why we should use samples of the same volume. Unfortunately, we had only a 80 GBytes disc, but we hope to test a 40 GBytes model as well in the near future. The High-End Disk WinMark uses templates calculated according to operation of real applications, and at the expense of their wide variety the effect is weaker.

NTFS lacks for such problem. For example, 40 and 80 GBytes discs are formatted with the same cluster size of 4 KBytes (it can be changed). The size of the identical table, MFT, is dynamic. It also positively tells upon the situation.

Now let's turn to the IOMeter test.

As usual, we use three patterns - Server, Database and Workstation with 5 versions of load (this parameter is marked on X-line). The vertical axis shows speed in MB/sec.

IOMeter Server

IOMeter Database

IOMeter Workstation

The performance of the Seagate U6 is not great here, first of all, because of the large access time which greatly affects the IOMeter results. The leaders are WD300AB and Fujitsu MPG3204AT due to a small access time. It's interesting that the graphs of the Seagate U6 and Maxtor DiamondMax VL40 are very similar to each other despite different access times. It's obvious that not everything depends on a graph of linear reading and access time: there are other hidden differences.

Conclusion

The new disc of the Seagate U6 series is undoubtedly a successful model. Thanks to a high record speed it's the first time when a 40 GBytes disc uses only one platter. Unfortunately, the increased record density affected an access time, but we hope it will be eliminated very soon.

An expected low price allows us to assume that the model will become very popular. High results of the ZDLabs Winbench place the new generation of Seagate disc into a leading group among 5400 RPM hard discs. Optimization of the internal software makes the highest contribution into it. The Seagate engineers managed to lift the performance when working with usual applications under Windows at the low cost. Low results of the IOMeter shouldn't be attached importance to since one will hardly use such disc in servers. At least it can be used in the IDE RAID to build an inexpensive system with a disc size up to 320 GBytes on one controller!


Write a comment below. No registration needed!


Article navigation:



blog comments powered by Disqus

  Most Popular Reviews More    RSS  

AMD Phenom II X4 955, Phenom II X4 960T, Phenom II X6 1075T, and Intel Pentium G2120, Core i3-3220, Core i5-3330 Processors

Comparing old, cheap solutions from AMD with new, budget offerings from Intel.
February 1, 2013 · Processor Roundups

Inno3D GeForce GTX 670 iChill, Inno3D GeForce GTX 660 Ti Graphics Cards

A couple of mid-range adapters with original cooling systems.
January 30, 2013 · Video cards: NVIDIA GPUs

Creative Sound Blaster X-Fi Surround 5.1

An external X-Fi solution in tests.
September 9, 2008 · Sound Cards

AMD FX-8350 Processor

The first worthwhile Piledriver CPU.
September 11, 2012 · Processors: AMD

Consumed Power, Energy Consumption: Ivy Bridge vs. Sandy Bridge

Trying out the new method.
September 18, 2012 · Processors: Intel
  Latest Reviews More    RSS  

i3DSpeed, September 2013

Retested all graphics cards with the new drivers.
Oct 18, 2013 · 3Digests

i3DSpeed, August 2013

Added new benchmarks: BioShock Infinite and Metro: Last Light.
Sep 06, 2013 · 3Digests

i3DSpeed, July 2013

Added the test results of NVIDIA GeForce GTX 760 and AMD Radeon HD 7730.
Aug 05, 2013 · 3Digests

Gainward GeForce GTX 650 Ti BOOST 2GB Golden Sample Graphics Card

An excellent hybrid of GeForce GTX 650 Ti and GeForce GTX 660.
Jun 24, 2013 · Video cards: NVIDIA GPUs

i3DSpeed, May 2013

Added the test results of NVIDIA GeForce GTX 770/780.
Jun 03, 2013 · 3Digests
  Latest News More    RSS  

Platform  ·  Video  ·  Multimedia  ·  Mobile  ·  Other  ||  About us & Privacy policy  ·  Twitter  ·  Facebook


Copyright © Byrds Research & Publishing, Ltd., 1997–2011. All rights reserved.