memory usage

[seaken64]

Have you tried MX-Fluxbox? It is an desktop overlay in MX. You still have your MX but the memory footprint is a lot lower. This is different than booting to antiX (which is another viable option for lower ram use).

There is a Sub-Forum on MX-Fluxbox in the forum directory. Check it out.

I have initial memory under 100MB on a MX-18 install. It is more on some of my MX-19 systems but still half of what is reported in XFCE.

Seaken64

[retired]

OK - so I didn't have a swap partition so I reinstalled and created a 2 gb swap - hope that resolves the issue - thanks for the support - regards to all

[Stevo]

Swap just keeps your system from freezing totally when it runs out of RAM. Having it slow down to a crawl as it uses swap is not much less frustrating, so you want to avoid using swap if at all possible.

[anticapitalista]

...

Between Antix/Bodhi & MX, there's not that much difference....

Sorry, but I can't agree.
Based on your own posted figures, there is a huge difference between these 3. antiX is below 200MB RAM, Bodhi is just above 200MB RAM and MX RAM usage is more than double antiX's at over 400MB RAM!

[Jerry3904]

That sounds right. Most users I have seen report 400+ for MX-19 with the more demanding Xfce 4.14, and about 200 with MX-Fluxbox.

[az2020]

...

Between Antix/Bodhi & MX, there's not that much difference....

Sorry, but I can't agree.
Based on your own posted figures, there is a huge difference between these 3. antiX is below 200MB RAM, Bodhi is just above 200MB RAM and MX RAM usage is more than double antiX's at over 400MB RAM!

Are we looking at the same thing? The table I posted above shows (after updating in the virtual box) Antix was 195mb, Bodhi was 214mb. [EDIT: Sorry, I realize you were emphasizing MX. I saw the "bodhi/antix" and stopped reading there. I think when I wrote what you quoted, 200mb didn't seem like that much compared to the full range of distros. But, I agree that could be substantial to someone on a low-resource machine. I'm thinking the range between MX & Antix/Bodhi are all considered lightweight. But, I agree MX is up there. And, that could make a difference for very resource-limited machines. I wish distros would publish a "how to make this as skinny as possible." Turning off samba, etc. could make a big difference, and make MX more comparable to Antix?]

I did a comparison April 2019 (posted in another thread). Antix 17.4.1 was 99mb. Bodhi (still 5.0.0 then) was 146mb.

It's interesting that Bodhi remained 5.0.0, but grew from 146mb to 214mb. I guess that's due to the system update bring more new stuff(?). Or, maybe there was a difference between Qemu (I used that last time) and VirtualBox (used this time).

I don't know how comparable the installs on real hardware are. To me, in terms of comparing distros, the vitrual environment seems more comparable about the distro itself. But, those installs don't reflect installation on real hardware with real drivers, doing real things. :) But, then the distros vary a lot depending on how well they support newer hardware, I suppose. For example, Bodhi's virtually the same whether installed in virtual or my relatively new Ryzen 3 3200u/Vega 3. Whereas other distros jump very high in mem use when installed on that laptop.

So, you have to weight those numbers for what they are. (I did document in the linked materials how I compared. Anyone can recreate that on other hardware.)

I'm going to this again this summer after the 20.04 stuff is out, and the LTS-based distros like Peppermint & Linux Lite are released. If there's something I can do different to shed more light, I will (assuming it easily fits into what I do. I don't want to spend a lot of time on it.). One thing I think I'm going to do differently is do the "free" command once every 5 minutes (for 15 minutes) and take the average. What I have been doing is waiting for it to settle down. But, that could be arbitrary. Sometimes they start high, and then stabilize lower. Sometimes they start low, then jump high. Sometimes they cycle up and down (and I try to take what looks like the average). Maybe I should take a "free" reading every 5 minutes (3 times, until 15 minutes elapse) and average that.

[anticapitalista]

I don't dispute your tests and the results of the tests (i find them very valuable), but I do dispute that particular conclusion you came to i.e. there is not much difference between antiX/bodhi and MX.

In your same test, there were other distros that showed significantly less RAM usage than MX e.g. Peppermint, Linux lite, Lubuntu, Sparky, Xubuntu and even KDE neon used less RAM than MX.

MX doesn't claim to be lightweight so it is no surprise that its RAM usage is higher than distros that do claim to be such as antiX and bodhi and possibly sparky.

[christophe]

@retired
FWIW, I found my MX-19 setup on my 2 GB laptop (Dell D420, 32-bit, frugal install) starts up at about 282 MB, and with Firefox (only this 1 tab open) is 530 MB -- 2 MX forum tabs brought it to 606 MB, with 68 KB swap used. I got here by tweaking it slightly with:

1. /etc/sysctl.conf added ("borrowed" from antiX):

Code: Select all

 vm.swappiness=10
vm.vfs_cache_pressure=50 

2. changed the kernel to the 4.9 antiX kernel (from the MX Package Installer).
3. disabled a couple of startup programs (from settings manager on top of whisker menu) that I KNEW I could live without, like clipit & blueman.
4. enabled zram for swap (from the wiki) -- and no hdd swap.

That made my old 32-bit 2 GB system very responsive & enjoyable for me to use. I'm not sure which part did what % of the improvement. I just threw everything at it that I thought would work.

(As a side note, my antiX-19 32-bit on the same computer boots up to 127 MB memory used.)
*** edit: Just noticed the 29.2 MB for my ROX panel launcher; that is what pushes my antiX over 100 MB.

[JayM]

On my 2009 Averatec N1130 netbook, 2gb DDR2-800 ram, 32-bit 1.6gb Intel Atom single-core CPU, with MX-19.1 and the antiX 4.9 kernel installed, using Fluxbox, with non-essential (to me) services like Bluetooth, samba, cups and saned disabled, no conky, using Falkon web browser instead of Firefox: 118MB idle memory use with just htop running in a terminal, 287MB in use with Falkon open to the forum (one browser tab open) plus htop.

Since OP never gave his Quick System Info it's impossible to offer him specific suggestions on how to reduce memory consumption on his computer. If he's interested in reducing memory use by reducing the number of service daemons that run in the background and using MX Fluxbox he can follow seaken64's instructions as I did.