On Monday 11 December 2006 18:57, Adam Talbot wrote:
Forget the nice way that suspens2 works, by only pulling the used RAM to the suspend image;
I always hated the silly way windows works on hibernation - at least you can use a list of used pages and just write them out to speed it up, eventually even use a fast compression algorithm for easy to compress pages or similar ideas.
At least Windows takes longer to boot from disk image (for large RAM as today) than normal boot, no gain there - would that be better for LinuxBios? I doubt it.
So what is so great to not boot from swap-image?
Wouldn't it be easier to patch filo to be able to boot the swap than to reprogram suspend2 ?
BTW - i thought suspend2 can also boot from a file instead of swap?
But i'am curious what timing one could expect when using suspend2 and LinuxBios, the time to graphical interface wouldn't be much longer than to console - just the time to swap in the X-server, i guess?