Did you do sudo swapoff -a? Try it with htop running, it's quite a cool sight.
What am I supposed to see with this way ?
I had 200 MB and watched it slowly drain away.
Hmm, i have not used swap for quite some time
Even when i had 2GB (i now have 8GB) or RAM, i still ran without a RAM partition.
I do not know when it is actually needed, i have a strong feeling that it is only in the most extreme of cases swap is needed today.
Often "video editing" is mentioned in this cases, but perhaps less often were the user actually has that need in real life workcases.
Though, i tend to use as little RAM as possible most of the time, i do things like: firefox 20+ tabs, skype, spacefm and minecraft (with 256x texture packs) simultaneously without sweat from linux.
The only times i have actually even "felt" my RAM being pressed, is when an application eats or leaks memory (like java for instance),
what has happened is a short freeze of the actual application (minecraft) and it being switfly killed by linux shortly thereafter.
Can it be so that some applications/system-thingie's actually tries using up and expanding further into to swap just because actually can?
- making it seem like you need so much for RAM space?
I have heard that a good system/application uses RAM when it can, because having it not used is well, useless.
Perhaps you could try removing the ram-partition (or maybe not mount it in fstab and then reboot) all together, and use the computer for a time period the way you usually do
- and see if it actually hinders you lacking one?
I don't know, these are just my personal experiences of the whole swap or no swap discussions
Last edited by PReP (2013-07-14 14:20:19)
. Main: Intel Core i5 6600k @ 4.4 Ghz, 16 GB DDR4 XMP, Gefore GTX 970 (Gainward Phantom) - Arch Linux 64-Bit
. Server: Intel Core i5 2500k @ 3.9 Ghz, 8 GB DDR2-XMP RAM @ 1600 Mhz, Geforce GTX 570 (Gainward Phantom) - Arch Linux 64-Bit
. Body: Estrogen @ 90%, Testestorone @ 10% (Not scientific just out-of-my-guesstimate-brain)
I find the system does use swap even though I have 8G of RAM. However, it is usually very small amounts of swap e.g. right now, I have 20M of swap used. Occasionally, the system will briefly become unresponsive but this is always when I am asking it to do a *lot* of things (especially if I am compiling code in the background). Mostly, it copes fine and it has only become unresponsive for a second or two even recently when I've been compiling kernels. (I think in that case I may have been compiling something else or updating the system simultaneously - usually the only sign of stress when compiling a kernel is that the temperature goes up, responsiveness is usually fine.)
I haven't changed the default swappiness from 60, though, because the existing use of swap really doesn't seem problematic.
@cfr, what --jobs (-j) setting do you use when compiling a kernel? I have MAKEFLAGS="-j6" in makepkg.conf, and like you , besides a slight increase in temp, I sitll can't seem to tell that it is working so hard.
-j5. I have an alternate config file for non-parallel builds but I use the default to compile the kernel. (There is something I compile I need non-parallel for though I forget what but it isn't the kernel.) Why?
I'm just curious, as I have always used -j6 for this machine. So I was just wondering what others use with similar hyper-threaded processors. I know that your machine has a Core i3.
I think I just went by the wiki recommendation to use (n+1) where n is the number of processors. I have 2 processors but they act like 4 so (4+1)=5. I've no idea if that advice is sound, though.
I think it is good enough advice. It will just ensure that the compilation really takes advantage of all the processing power that is available to it. If you use -j1 you will see that only one thread of your processor gets used, so presumably 4 *should* use all four logical cores, but 5 will certainly make sure of it. I use 6 just to make sure that I am sure that I am going to be building as fast as possible. But when I started using -j6 with this machine, I was just kind of surprised that there was no noticeable slowdown on my machine, whereas with my old MacBook Core2Duo 1.83GHz, using -j3 would cause significant lag.