Re: nifty swapspace trick

From: Arthur Corliss <acorliss@nevaeh-linux.org>
Date: Mon Oct 29 2007 - 14:39:13 AKDT

On Mon, 29 Oct 2007, Larry Collier wrote:

> Arthur,
>
> I agree with everything you've written so far. Given no external
> limitations, more real ram is far better than anything else you can do.
>
> But what is the solution for the old laptop, of which I have two, that is too
> slow due to limited ram, but is also not upgradeable in any way, since the
> standard practice back then was to put the ram on some little dingus board
> buried under three others requiring two special screw drivers to get to.
>
> This is a good solution to get a few more miles out of them before they go to
> the recyclers.

Tough call. An old laptop probably has a horrible drive running @ 4,500
RPM, but it also probably has USB 1.0. Neither option is pretty. Now, if
you are using a PCMCIA slot with a flash media reader, you may be right.
For entertainment's sake, I'd love to see some benchmarks comparing the
flash to the internal drive. :-)

> Going off on a small tangent, what do you think about the plans of I think
> Samsung to replace the drive with flash in laptops. Have they improved the
> write cycle limitations to the point that this becomes viable?

That actually becomes a question of drivers and software. They can solve
some problems in software (the embedded Linux kernel community is working on
drivers to spread writes across all blocks to lengthen the life of the flash
media). The problem there is that in order for this to be viable you need
*lots* of slack space. If you fill up your drive with a lot of software,
mp3s, etc., your swap & temp space will still kill you because they'll be
writing out to a very small number of blocks.

Even so, I think it's only viable for those that are locked into planned
obsolescence. If you know you're going to upgrade every three years, they
might very well be able to do flash drives in a manner that will never
impact you.

         --Arthur Corliss
           Live Free or Die
---------
To unsubscribe, send email to <aklug-request@aklug.org>
with 'unsubscribe' in the message body.
Received on Mon Oct 29 14:39:30 2007

This archive was generated by hypermail 2.1.8 : Mon Oct 29 2007 - 14:39:30 AKDT