[aclug-L] Re: Linux as an alternative to a Windows desktop

[Jeff Vian <jvian10@xxxxxxxxxxx>]

Nate Bargmann wrote:

>* bruce <bbales@xxxxxxx> [2002 Dec 16 21:05 -0600]:
>  
>
>>The result was that the heads had to move less on the outer 
>>tracks and the data came off faster (more bits per revolution). 
>>I suspect this scheme is still used - you'd have to ask a disk 
>>designer.
>>    
>>
>
>That jogs my memory and I seem to recall that this technique is what's
>being used on modern drives.  This would take into account that a
>certain amount of magnetic material is required for X amount of data
>storage and there is more per track at the outer edge.
>  
>
IIRC you are correct (the sector-per-cylinder numbers are at best an 
average and more likely highly underrated). IIRC the C-H-S number of 63 
sectors per cylinder has not changed in several years, and drive 
capacity is increasing exponentially.
On the older slower drives the data was actually contiguous in 
sequential sectors. On todays higher speed drives I believe a standard 
interleave of 3:1 is used, Thus contiguous sectors in a track are 
actually every 3rd sector physically. This allows for reads and writes 
with data transfers in the buffers during the skips of sectors -- faster 
through-put overall.

With these updates in drive performance, the major slowdowns are the 
requirements for head seeks. This becomes a factor in fragmented files, 
or repetitive seeks to data located in  different areas of the disk.

> 
>  
>
>>If it is, it could have a big impact on where the best place to 
>>put data really is.
>>    
>>
>
>Would it, or does it tend to even things out leaving the best place to
>put data a matter of keeping frequently read/write data together?
>
>- Nate >>
>
>  
>


-- This is the discussion@xxxxxxxxx list.  To unsubscribe,
visit http://www.complete.org/cgi-bin/listargate-aclug.cgi