On Tuesday 19 August 2003 23.52, Scharf Yuval wrote:
> Nicolas, If you are reffering to hdb you are wrong because there is no
> device there.
>
> So, If I sum up your answers then my HD is capable of 30MB/s.
> The UDMA(100) is a sales persons trick.
> My bus is 33Mhz and it is not holding back my HD.
>
> So, If I like to buy a new motherboard and a new HD, what should I buy?
> Assume that I would like to get the best a person might expect from a home
> computer.
>
Try going to a 2.6 kernel.  I got around 20% boost in through put by doing 
this.  I think there is also some expermental IDE stuff you can enable in 2.4 
which could give you a boost.  I played with it but that was a long time ago.  
This what I get with 2.6-test3 kernel, AMD SMP and a WD drive.
Timing buffered disk reads:  114 MB in  3.04 seconds =  37.54 MB/sec
I use to get figures around the same as you with 2.4.  I get this in the same 
machine using a Promise IDE raid controller.
Timing buffered disk reads:  138 MB in  3.03 seconds =  45.52 MB/sec
and as I raid 0 array (kernel raid 0)
/dev/md0:
 Timing buffer-cache reads:   1036 MB in  2.00 seconds = 516.79 MB/sec
 Timing buffered disk reads:  190 MB in  3.01 seconds =  63.17 MB/sec
On my other machine with an nforce2 MB and seagate drive with a 2.4 kernel 
(gs-sources)
media root # hdparm -Tt /dev/hda

/dev/hda:
 Timing buffer-cache reads:   1456 MB in  2.00 seconds = 728.00 MB/sec
 Timing buffered disk reads:  144 MB in  3.04 seconds =  47.37 MB/sec

So MB/IDE Host controller and kernel all affect the results, probabley more 
that actual recent HD type.

tony


-- 
Contract ASIC and FPGA design.
Telephone +46 702 894 667
http://pgp.mit.edu:11371/pks/lookup?op=vindex&search=0x633E2623



--
[EMAIL PROTECTED] mailing list

Reply via email to