Fujitsu company is traditionally dealing mainly with OEM-partners and delivers only remaining products via distributor channels. That is why it was always difficult to get devices of this company. We still can't get the latest model of Fujitsu with 7200 rpm spindle speed - MPG3xxxAH-E.

Today we are going to consider two models - MPG3204AT of 20.4 GBytes capacity and 5400 rpm spindle rotation speed and MPG3102AH of 10.2 GBytes and 7200 rpm. The discs have Ultra ATA/100 interface. Unfortunately, we failed to get discs of the same size. The characteristics are shown in the table.

	Fujitsu MPG3xxxAH	Fujitsu MPG3xxxAT
Capacity	10.2/20.4 GBytes	10.24/20.4/30.7/40.98 GBytes
Spindle rotation speed	7200 RPM	5400 RPM
Buffer size	2 MBytes	512 KBytes/512 KBytes/2 MBytes/2 MBytes
Number of heads	2/4	1/2/3/4
Number of discs	1/2	1/1/2/2
Average Latency time	4.17 ms	5.56 ms
Average reading seek time	8.5 ms	9.5 ms
Average track to track seek time	0.8 ms	1.0 ms
Average full track seek time	16 ms	17 ms

Whereas the MPG3xxxAT discs correspond to the up-to-date level: 20 GBytes per platter, heads with gigantic magneto-resistive effect of the 6th generation, the elder brother doesn't look so solid - there are only 10 GBytes per platter. Let's turn to the results.

Tests

At our site we are conducting the benchmarks of the hard discs according to the special method. The results are following.

Ziff-Davis WinBench 99

The results are expectable. Since the speed of linear reading directly depends on the record density, 10 GBytes per platter doesn't allow to count on something better. Nevertheless the graphs of reading are ideal.

MPG3102AH - FAT32

MPG3102AH - NTFS

The discs with 5400 rpm speed have better results.

The Fujitsu is just a bit behind the leader - Seagate U5, but on the whole it performs quite well. However, it can't boast of its graphs.

MPG3204AT - FAT32

MPG3204AT - NTFS

As a matter of fact, the "shivering" of the graph at the first stage is typical to all discs with 20 GBytes per platter capacity because of high data density. The Seagate U5 and Quantum lct20 make an exception due to the low spindle rotation speed. And "shivering" at the end doesn't take place often. It is interesting that the linear reading speed of the disc with 5400 rpm speed and 20 GBytes per platter density is higher than that of the disc which rotates quicker but has 10 GBytes per platter. But there is nothing strange, and 7200 rpm discs have their own advantages, for example, lower access time of the Fujitsu MPG3102AH.

Now about the access time.

This parameter doesn't depend on the record density, that is why the Fujitsu has good results here. The benchmark time is just a little higher than the specified one (12.67 ms).

Here the Fujitsu performs also quite well - a comfort middle with a slight exceeding of the specified value (15.06 ms).

Now it's high time to check the operation with standard applications.

The results are again logical. Whereas in applications working with small-size files (for example, MS Frontpage) the MPG3xxxAH disc is more or less competitive, in applications which require reading of large files the disc is affected by the small record density.

It has a strong second position in the High-End Disk WinMark test for both file systems. Besides, it outdoes the Samsung nearly in all applications. The Fujitsu MPG3104AT doesn't like very much the Adobe programs (Photoshop and Premiere). But as to the sound, there is nothing to complain about - it is a leader in Sound Forge 4.0. The MS Visual C++ suits the disc only with the NTFS in use. In the Business Disk WinMark, though, the results are worse.

The complete benchmark results are shown in the table.

Intel IOMeter

The test is still puzzling. It is clear that for realizing the results it is necessary to collect some statistic data, what I am doing right now. But it is still vague how to introduce the results. If you wish you can view the complete table of results. Since it is difficult to conceive what is behind the figures, I have decided to choose one parameter - a data volume transferred at one second (MBps), and to draw graphs of this parameter depending on the load for each access pattern.

So, first of all, the 7200 rpm discs

It is clear that the record density doesn't influence the results of this test. However, the result is directly dependent on the disc access time.

And what about the discs with a spindle speed of 5400 rpm and lower?

It is proved that at low loads the access time plays the most important role - in all other respects the lct20 is significantly inferior to the U5. With the growth of the load the firmware starts to influence the final results in quite a curious way.

Conclusion

In my conclusion words I will be guided by the results of the Ziff-Davis WinBench test.

The Fujitsu MPG3204AT is, undoubtedly, a good disc. In all parameters it is just a bit behind the leader, but nevertheless, the device will be a perfect purchase.

The Fujitsu MPG3102AH is a disc of the previous generation. It's a high quality product (look, for example, at the graphs of reading), but it is still inferior to the competitors in speed.

Write a comment below. No registration needed!