As I stated the article I was asked to write was on MS Windows O/S and
has not the slightest thing to do with Linux and I expressed my joy of
not having to deal with it any more.
Scott

M Harris wrote:
> On Wednesday 09 May 2007 14:20, Registration Account wrote:
>   
>> Personally IF the PC has the max amount of physical RAM installed AND
>> still requires a Disk I/O to simulate more RAM then we really need to
>> think about the fundamental operation of the O/S.  How much RAM can it
>> directly address and how much RAM does it need to juggle in and then
>> does it requires a page file as well - Personally then its time to
>> re-write the O/S in its memory management and its directly addressable
>> RAM to process and the amount of physical RAM used by the memory manager.
>>     
>       The above paragraph is an over simplification, and not entirely correct 
> in 
> general, and not correct at all in terms of Linux.  There are legitimate uses 
> for a swap file... and many linux installations make use of several large 
> swap files on each machine!  My current desk machine has a swap file of 1024M 
> and real ram of 512M.  I will generally set my swap file to twice the size of 
> main store--- but there is no such fixed rule. Having said that--- most of 
> the time my machine *never* swaps. I've been watching my new system monitor 
> (thanks Randall) now most of the afternoon and my machine has not swapped... 
> not once.  As I have used the machine (mostly mail, small compiles, web 
> research) I have noticed the kernel adjusting my cache and buffer sizes and 
> moving along quite happily. This in contrast to my old W2000 and NT machines 
> that would constantly *thrash* after moderate use... this means many page 
> faults and lots of wasteful disk I/O like you talk about in your article. The 
> windoze platform has lousy memory management and even worse file system 
> management. The Linux kernel has *none* of these problems.  This is no 
> exaggeration--- memory management on Linux is comparable to memory management 
> on MVS/XA, VM, SysV, Sys38, AS400, name it... seriously.
>
>   
>> Page file addressing should be a last resort by the O/S. Adding to an
>> overworked disk I/O will slow things down ultimately, however the
>> application will never fall over and you will never see the old "out of
>> memory" error response which is the only advantage of such an arrangement.
>>     
>       (see above) Paging is used by the Linux kernel on a needs basis-- 
> absolutely. 
> And, sometimes those needs are very real and quite legitimate. The more you 
> play around with the Linux kernel the more comfortable you will become of 
> course, but rest assured--- you're in good hands.  Its easy to create stress 
> scenarios for your machine that will tax memory and force paging... check it 
> out. 
>
>
>
>   

Attachment: smime.p7s
Description: S/MIME Cryptographic Signature

Reply via email to