iXBT Labs - Computer Hardware in Detail

Platform

Video

Multimedia

Mobile

Other

Fall 2001: SCSI HDD (10K and 15K RPM) Roundup

November 30, 2001



The last large roundup of SCSI discs was published last summer. Unfortunately, not all manufacturers launch often new models of the high-end class. We waited a year for new products from the giants. And as we can't wait any longer we think it is high time to reestimate the leaders in these sphere.

First we planned to consider leaders in the 10K RPM class and one model of the 15K RPM class from Seagate. But just a little time ago we received a good competitor against the Seagate drive. It is a hard drive from the Fujitsu MAM series. Unluckily, a similar device from IBM will appear by the New Year only. That is why the 15K RPM class will be represented by 2 drives.

First come the specifications:

  Fujitsu
MAN3xxx
Fujitsu MAM3xxx IBM
Ultrastar 73LZX
IBM
Ultrastar 36LZX
Maxtor
Atlas 10K III
Seagate
Cheetah 36XL
Seagate
Cheetah 73LP
Seagate
Cheetah 36ES
Seagate
Cheetah X15 36LP
Capacities (GB) 18.4
36.7
73.5
18.4
36.7
9.1
18.3
36.7
73.4
9.1
18.3
36.7
18.4
36.7
73.4
9.2
18.4
36.7
73.4
18.4
36.9
18.4
36.7
Heads 2/4/8 4/8 2/3/6/12 3/6/12 2/4/8 2/4/8
8
2/4 4/8
GB/Head
9.2
4.6
6.1
3.0
9.2
4.6
9.2
9.2
4.6
RPM
10025
15000
10000
10000
10000
10022
10041
10000
15000
Seek Time Average (ms)
4.5
3.5
4.9
4.9
4.5
5.4
4.7
5.2
3.6
Average Latency (ms) -
2.00
3.00
2.99
3.00
2.99
2.99
2.99
2.00
Internal Data Rate (MB/sec) 416-651 320-710 376-697 280-452 350-622 320-490 399-671 520-682 522-709
Sustained Throughput (MB/sec) - - 30-58 21-36 33-55 31-50 38-64 49-63 51-69
Buffer Size (MB)
8
8
4
4
8
4
4
4
8
Interface Ultra160 Ultra160 Ultra160
Ultra320
Ultra160 Ultra160
Ultra320
Ultra160 Ultra160 Ultra160
Ultra320
Ultra160
Ultra320
Height (inches)
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
Sound Power (Bels, Idle) 3.6/3.6/3.9 3.6/3.9 3.7/3.4/3.4/3.4 - 3.1/3.4/3.6
3.8
3.6
3.2
3.5
Typical Power Draw (W, Idle) 7.8/8.4/9.2 9.5/11.5 6.7/6.7/7.4/9.5 8.5/9.7/12.9
10
7.5/8/9.6
10.2
8.5
10.5/11.5

It is quite difficult to compare hard drives bearing in mind only their specs as every company has its own methods for measuring different parameters. And the only comparable values are dimensions, volume and weight :)

As you can see, Seagate has the greatest number of models. A figure in their names means the maximum capacity in the series and letters mean generations: without letter, LP, XL, ES. Earlier, speed characteristics of the series of different capacity were almost equal, for example, 18LX ~ 36 LP ~ 73, but starting from 36XL/36ES/73LP it isn't so. And since the 73LP series much differs from the 36XL one we have added one model of this series to the 36 GBytes discs.

The Cheetah X15 36LP discs belong to the second generation of 15K RPM HDDs from Seagate. And the Fujitsu drive is the first model with a spindle speed equal to 15K.

We failed to find drives from the last 10K RPM series of IBM - Ultrastar 73LZX. We are also expecting a model from the 15K RPM class from this company by the year end. At the first opportunity we will update the review with their results.

Tests

Test platform:

  • processor: Intel Pentium III 800EB
  • mainboard: ASUS CUSL2-LS (i815, Adaptec 7899 Ultra160)
  • memory: 256 MBytes PC133 SDRAM
  • video: integrated (i815)
  • system disc: IBM DTLA-307015
  • OS: Windows 2000 Pro SP2

Hard Drives:

  • Fujitsu MAN3xxx: MAN3367MC
  • Fujitsu MAM3xxx: MAM3367MC
  • IBM Ultrastar 36LZX: DDYS-T36950
  • Maxtor Atlas 10K III: KW36J11
  • Seagate Cheetah 36XL: ST336705LW
  • Seagate Cheetah 36ES: ST336706LW
  • Seagate Cheetah 73LP: ST373405LW
  • Seagate Cheetah X15 36LP: ST336752LW

Photos of the participants:

MAN3367MC

MAM3367MC

DDYS-T36950

KW36J11

ST336705LW

ST336706LW

ST373405LW

ST336752LW

Only Fujitsu models are similar in appearance. For the first time for Seagate its drive, namely Seagate Cheetah 36ES drive, uses a smaller PCB, and an engine is not seen from the back side.

In the Fujitsu and Maxtor devices the drive mechanism doesn't come out to the back side, like in other models. The Cheetah 73LP drive has almost the same PCB as in the first Cheetah X15 series. But the second generation - Cheetah X15 36LP - has a new motor and a completely redesigned layout (only the flash memory under the firmware hasn't changed its size).

Tests:

  • ZDLabs Winbench 99 1.2
  • HDTach 2.61
  • Intel IOMeter

For the Winbench the discs had one partition of FAT32 and NTFS. For all other programs any information on partitions was deleted. The Winbench and HDTach tests were carried out 3 times, and the results were averaged. The operating time of one pattern at each load in the IOMeter was set to 10 min. It should be noted that the figures on the diagrams are rounded, though for plotting we used more precise ones.

Access time

All three tests were used to measure an access time. For the IOMeter we used the "Average Response Time" as a result for the pattern of 100% random reading with "Queue Depth" equal to 1.

On the whole, the results obtained by different programs are similar, and the Cheetah 73LP and Cheetah X15 36LP drives have scored identical results in the WinBench 99 and HDTach.

The aging Ultrastar 36LZX is still able to fight against the latest generation of the Seagate Cheetah. But it is too weak for a struggle against the latest models from Fujitsu and Maxtor. The latter ones show almost equal results in these tests. The 15K RPM drives are, of course, leading. By the way, as compared to the first generation of 15K RPM discs from Seagate the Cheetah X15 36LP improved its average access by 12%! The new-comer from Fujitsu approaches it.

The Cheetah 36ES and Cheetah 73LP go on a par, yet despite Seagate announcements, different designs and different generations.

Linear reading speed

The HD Tach gives average results:

These results were expected as we knew the recording density data. The discs with the density of 9.2 GBytes per side have the closest results. And again the Cheetah 73LP is much closer to the 36ES than to the 36XL.

The discs of the RPM 15K take the lead again, though the breakaway from the 10K RPM group is not so great: 8% in reading and 24% in recording.

The detailed graphs over the whole disc surface are obtained with the Winbench program (percentage of a disc size is plotted off the X axis):

The IBM comes the last preceded by the Cheetah 36XL with a low density of 4.6 GBytes per side. The leader is, of course, the Cheetah X15 36ES. It is interesting that the other drives have very similar results. If we take into account almost equal access times of the Fujitsu and Maxtor, the integer tests will reveal the gain due to a firmware optimization. As this past of a disc is hidden from a user, there can be a lot of interesting algorithms.

In the 15K class the Fujitsu lags behind the Seagate by 6% at the beginning, though by the end the gap gets almost zero.

Winbench Integer Performance

Despite quite good results in the previous tests the Fujitsu MAN3367MC failed to catch up with the group of Atlas 10k III and Seagate drives. The Cheetah 36ES goes ahead in the 10K RPM class while the Fujitsu MAM3367MC falls into the last position. Despite a little access time it lost in the FAT32 even to the Cheetah 36ES.

Here the Fujitsu disc has got closer to the Cheetah 36ES and 73LP, while the Maxtor Atlas 10K III has broken away from its 10K RPM contestants from Seagate and is doing its best to beat the 15K RPM drives. In the FAT32 it easily outscored the Fujitsu and almost caught up with the Cheetah X15 36LP.

Intel IOMeter

The IOMeter tests are carried out with the Fileserver, Workstation and Database patterns. The load parameter - "# of Outstanding I/Os" is plotted off the X axis.

Here we got a new leader - Fujitsu MAN3367FC. Earlier this drive yielded to the Atlas 10k III, and now it takes a 3-5% lead in the Workstation test and even 10-12% in the File Server and at a heavy load. The last two positions are taken by the Cheetah 36XL and Ultrastar 36LZX, though the gap from the other models is quite small.

The 15K RPM drives are more efficient than its predecessors: the difference is 3045%. It is mainly caused by a less access time which is a determining factor in the IOMeter tests. The Fujitsu outdoes the Cheetah X15 36LP by 37% at a low load, and this is an excellent result for a new-comer.

Temperature and noise

The temperature was measured by an external digital device on the side panel of the hard drive after 30- min. operation of the IOMeter test with the random reading pattern.

The Fujitsu MAN3367MC and Maxtor Atals 10K III shine in this test as well. The new model in the 15K RPM class - MAM3367MC, unfortunately, yields to the ST336752LW.

Unluckily, we do not have an opportunity to measure correctly a noise level, that is why our estimation is subjective to a great extent. Besides, usually such hard drives are a part of arrays which obviously are not quiet. On the whole, you may trust to the official data. Each new generation of drives is quieter than the previous one. All modern drives of the 10/15K RPM class are not too quiet but in comparison to the first Seagate Cheetah devices they make much less noise. We think that the Atlas 10K III is the quietest in the 10K RPM class. The Fujitsu loses to its competitor with a spindle speed of 15K.

Conclusion

Although Seagate is leading an aggressive policy other companies do not give up. The long-awaited Maxtor Atlas 10K III performed quite well, especially in the integrated ZDLabs tests. This disc was announced in Q1 2001; it belongs to the third generation of 10K RPM discs from Quantum. It should be noted that reorganization of Quantum and Maxtor caused delays in delivery of this model, but it still successfully competes against the latest products from Seagate.

Fujitsu keeps on producing high-quality products. The MAN3xxx series is the fifth generation of 10K RPM drives of this company. Its first drive in the 15K RPM class is quite competitive as compared with the G2 of 15K drives from Seagate, a leader in this field. Both drives from Fujitsu take the leading positions in this class in the IOMeter tests.

Seagate is a confident leader in technologies. Among its products there is the fastest disc of 15K RPM - Cheetah X15 36LP. It is the second generation of X15 discs which beats the first one by 12% in the access time and by 50%(!) in linear reading speed. In the 10K RPM class the performance is gradually increasing as well: the Cheetah 36ES and Cheetah 73LP models of the G5 and G6 in the Cheetah series outscore the first 15K RPM models from Seagate in the Winbench test.

And now we are waiting for IBM to launch its 10K RPM Ultrastar 73LZX model and the second competitor against the Cheetah X15 - Ultrastar 36Z15.

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.