How to Hard disk write speed up?

[Tested Object]

Main Board: M1 (model:WD 160GB WD1600AAJS), M2(model: GA-MA78GM-US2H AM3)

Hard Disk: H1 (model:WD 400GB WD4000AAKS-00A7B0), H2(model: ESTAR ST A880G+ HD)

Lo-Speed: 20 - 30 MByte/s

Hi-Speed: 60 - 70 MByte/s

[Test Manner]

Test Progress: write test on large volume data by random offset not in sequence

Test Manner: processed in Linux App independently made by ourselves (asynchronous I/O)

[Test Process]

  1. M1-H1: Lo-Speed velocity comes out

  2. M2-H2: Hi-Speed velocity comes out

(Paragraph 1 and 2 above are under the initialized condition as originally set up)

  1. M1-H2: Hi-Speed velocity comes out

  2. M2-H1: Hi-Speed velocity comes out

(From the initialized settings of paragraph 1 and 2, cross settings of H1, H2 with M2, M1 are made)

  1. M1-H1: Hi-Speed velocity comes out

  2. M2-H2: Hi-Speed velocity comes out

(Settings are same as paragraph 1 and 2, but the velocity comes out as Hi-Speed)

[Test Result]

Many PCs of the same model have been tested to find that;

When H1 being combined with M2, does not go through the write test process, but goes through the POST process only,

All the combinations of M1-H1, M2-H1 produce the Hi-Speed velocity.

In other words, if the process of paragraph 4(combining H1 with M2) is taken at a single time only, any combination produces the Hi-Speed velocity.

[Test Analysis]

From the above tests, followings are assumed;

  • In the POST process, M2 is likely to have set up a certain value(related to speed) on to HDD

  • M1 is likely to have done nothing with settings or checking


  1. What would make the above results?

  2. Is the POST process in the combination of M2-H1 necessarily to be taken in order to advance the velocity of HDD by M1-H1?

  3. Are there any utilities for advancing the HDD speed by M1-H1?

  4. Are there any particular settings for BIOS?

  5. What would be the technological issues to advise us?


  1. We constantly have the same results as above with M1-H1 combination (200 units) M2-H2(100 units) in our possession.

  2. Still the same result from our tried tests with changed BIOS settings all relevant to HDD

  3. No differences are found between the two HDD even when we checked with the setting values of DMA, WCE, SCE … etc from H1 and H2 in use of the utilities as like hdparm, sdparm etc.


We look forward to your attentive, positive and affirmative reply while above combination is being used not only by us but also by many of our customers who are with an embarrassment as above.

Sorry, I don’t understand exactly what you are doing.

My first observation is that you have mixed up the Ms and Hs in the opening [Tested Object] section.

AIUI, in [Test Manner] you are saying that you are subjecting each motherboard and HDD combination to a random write test under Linux, using a large data set, beginning at random, non-seqential LBAs.

IIUC, in [Test Process] you are saying that you begin testing each drive as it is configured from the factory, using the M-H combinations in 1 & 2. This gives low and high speed results, as indicated.

In 3 & 4, are you saying that you perform the same tests again, by swapping the drives between motherboards? Are you powering down between the tests, or are you hot-swapping the drives?

Once again, in 5 & 6, are you powering down, or hot swapping? You appear to be suggesting that some setting has been altered on the drives.

Your [Test Result] section is difficult for me to comprehend. Are you saying that, if you take a virgin H1 drive and allow it to go through a BIOS POST on M2, without a subsequent write test, then H1 inherits some setting that then enables it to perform at high speed on either motherboard?

AIUI, in [Test Analysis] you are saying that M2’s BIOS reconfigures the drives, but M1’s BIOS does nothing. If you are powering down before swapping drives, then this would mean that these changes are non-volatile. I would capture the 512-byte information block being returned by each drive in response to an ATA Identify Device command, and then compare these data before and after each test.

An answer Thank you.

Was a sequential write test speeds were randomly will get the same results.

All power kkeunsangtaeeseo hadeudiseukeuman each other was tested by changing.

No such thing as any other settings that do not carry out the test has been changinghadeudiseukeuman servo.

I am also in the process of M2 POST board hard disk, set the value of any non-volatile, why not just make guesses.

I’m sorry, that didn’t translate very well (some phrases look like Hangul to me).

If I understand correctly, you are saying that you get the same results for sequential and random write tests.

You also appear to be saying that the hard drives were powered down, not hot swapped, and that you have not altered any settings yourself.

I don’t understand your last sentence. :frowning:

Could you please correct the following errors in your [Tested Object] section?

M1 (model:WD 160GB WD1600AAJS)
H2 (model: ESTAR ST A880G+ HD)

Could we see the Identify Device data from each drive?

By the way, you have posted to a user forum, not WD Tech Support. We’re all just users like yourself.

I notice that your Gigabyte motherboard has an Xpress Recovery2 BIOS. I don’t know whether your model is affected, but Gigabyte motherboards with an Xpress Recovery BIOS will automatically write a backup copy of the BIOS at the end of the drive and hide it in a HPA. The drive’s capacity will then be slightly reduced. I wonder if the drive’s settings are also modified at this time. As I understand it, only the first drive in the boot order is affected. You may like to see whether the POST modifies the settings of a virgin drive if it is installed as a secondary drive, or if it is connected via eSATA.

Gigabyte GA-MA78GM-US2H (rev. 2.0)

Support for @BIOS
Support for Q-Flash
Support for Xpress BIOS Rescue
Support for Xpress Install
Support for Xpress Recovery2

South Bridge:

1 x IDE connector supporting ATA-133/100/66/33 and up to 2 IDE devices
5 x SATA 3Gb/s connectors supporting up to 5 SATA 3Gb/s devices
1 x eSATA 3Gb/s port on the back panel supporting up to 1 SATA 3Gb/s device
Support for SATA RAID 0, RAID 1, RAID 10 and JBOD