Optmal Buffer Size while using FileRead?

[tip]

Hi all,

I'm a newbie and if this is a silly question please be patient with me ...

I was wondering how the buffer effects softwares... For ex. when we try to calculate MD5 hash of a large file we divide it to smaller pieces during FileRead.

for $i to ceiling($Filesize/0x20000)
      FileRead($OpenedFilesHandle,0x20000)
next

I'm not sure that "0x20000" can be called "Buffer" though I was wondering what effect it has on the script.

I tried to set it a higher value like 0x80000 but I couldn't noticed any difference while it was running.

So if you can help me figuring it out I'd be very content.

Thanks in advance

Edited April 15, 2009 by tip

[tip]

No one??

[Authenticity]

Read the CreateFile API function description and remarks you'll have a clue what buffering is good for and in which cases. Generally, if you're going to parse (copy) big files, more than a few mbs, you'll see a performance issues and maybe allocation failures if the file is above a few tens of mbs. It'll undermine your program stability as well as the system available memory. If you're going to develop things as "minimum required 512 MB RAM and recommended 2 GB RAM" you'll need to adopt buffering and determine when to use it, and when buffering is not required.

Edited April 16, 2009 by Authenticity

[tip]

So that "0x20000" has nothing to do with "buffer" right? You are saying buffer is something completely different.. Because when I calculate MD5 Hash with that line it consumes only 6MB of ram...