-=O0~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~O0=-
"He took his vorpal sword in hand:
Long time the manxome foe he sought -
So rested he by the Tumtum tree.
And stood awhile in thought."
[L.Carrol "Jabberwacky"]
On Tue, 22 Jan 2002, Maxim Kryachko wrote:
> Hi list.
>
> I run HD performance tests on Linux (RH7.1, kernel 2.4.18-pre3-ac2). The
> test is quite simple: first, I create a 2 Gb file filled with 1-s, then
> I read it and measure the time needed for both read and write
> operations. Tests run on 2 kinds of HDs: Barracuda ATA IV (ST380021A) -
> 80Gb, and IBM Deskstar DTLA-307076 - 75Gb.
> On Barracuda: It turned that writing of 2Gb file takes between 72 and 74
> sec, which is approximately equals to 27Mg/sec. Reading of the same file
> would take between 122 and 133 sec, which corresponds to read
> performance between 15Mb to 16.3Mb per second.
> On IBM: Both write and read performance would give about 27-29 Mb/sec.
> Both tests run on same kernel and same OS version. DMA enabled. 32-bit
> transfer enabled.
>
> Does anyone know what could be a problem? Why Seagate is so slow? Is it
> HD problem, or maybe some kernel tuning would help?
> Seagate support people say they don't know anything about Linux-specific
> issues.
>
i don't think this has anything to do with linux ...[though you run linux]
i suggest you to be fair to both disks:
the IDE controller position matters.
if your barracuda is not single on its IDE bus, then you get less speed.
numbers show difference of about 50% ... so i think you should check my
assumption.
> Thanks
>
> Max.
>
You welcome.
Max.
=================================================================
To unsubscribe, send mail to [EMAIL PROTECTED] with
the word "unsubscribe" in the message body, e.g., run the command
echo unsubscribe | mail [EMAIL PROTECTED]
