Clear the computer's memory?

[Tony Arnold]

Vince,

Vincent Trouilliez wrote:

> Oops, sorry for misunderstanding you... being a Frog, I will try and
> hide behind our reputation of being kinda disabled as far as languages
> go... ;-)

>From what I've seen of your postings, your English is certainly far
superior to any attempts I might make in French! And the Brits are
renowned for not speaking other languages. We just assume everyone else
should speak English:-)

>>And again, it Linux's paging system that causes the return to memory.
> 
> 
> Okay okay, I understand (I think ! ;-).
> So basically, Linux uses the standard swap mechanism to resume, hence
> the sluggishness. Okay makes sense. 
> The problem is that I assumed that there was actually a specific/purpose
> designed mechanism involved in the resuming phase, and that if the swap
> partition was used, it had nothing to do with swapping, but rather that
> it was the only partition that the system could safely use without
> risking to overwrite system or user files.
> So in my mind, the resume process could as well use any partition as
> long as it had enough space left to dump the RAM contents in it. 
> My assumption was wrong then....
> 
> Please, tell me that this time I got it right ?! ;o)

Well, that's as I understand it!

> That being said, hopefully the mechanism will be improve/sophisticated,
> to overcome this problem. Maybe do something like my original
> assumption: use any partition with enough space, dump the contents of
> the RAM, and reload it back into RAM in the exact same state that it was
> before hibernating. Basically, kind of making a binary snapshot of the
> RAM, put it to disk (swap partition or anywhere else), then copy that
> "image" back into RAM. A bit like cloning a hard disk with 'dd' ;-)
> This way, the system will be restored 100% like it was, hence if the
> system was responsive before hibernating, it will be as responsive after
> the resume. And if no swap was used before, not swap will be used after.
> Should speed up the resume phase as well, since I usually have only
> 200MB of RAM used, which tales only 4 or 5 seconds to read from a hard
> disk.
> 
> Well, something like that... :-)

An interesting idea that works if memory usage is reasonable low,
however, as memory usage goes up there must come a point where it is
overall better to bring stuff back into memory as and when it is needed
rather than to bring everything back in one go. If 200MB takes 4
seconds, then 1GB would take roughly 20 seconds. Maybe that's still OK,
but rather than wait 20 seconds you could be back running an
applications much quicker than that.

I do agree that hibernation seem sto be a sluggish process. My laptop
seems to take quite a while to go into hibernation mode, and then some
time to recover. I've always felt it could be quicker!

> Vince, day dreaming...

I have no problem that! Keep it up:-)

Regards,
Tony.
-- 
Tony Arnold, IT Security Coordinator, University of Manchester,
IT Services Division, Kilburn Building, Oxford Road, Manchester M13 9PL.
T: +44 (0)161 275 6093, F: +44 (0)870 136 1004, M: +44 (0)773 330 0039
E: tony.arnold at manchester.ac.uk, H: http://www.man.ac.uk/Tony.Arnold