Thumb2 vs non-thumb memory usage comarisions
 

During today's IRC talk, freemangordon expressed wish for proper comparison of thumb2 memory benefits. By a chance, I had non-thumb device at hand, and per vi_'s idea and request, here is thread for comparing memory usage and plotting conclusions.

For a start, here is my non-thumb output, took from device sitting idle for a few hours, without anything (except terminal, for executing measuring commands) opened:

Top:
cat /proc/swaps:
free:
/Estel

Re: Thumb2 vs non-thumb memory usage comarisions
 

may also help if you describe a few more things
1. your uptime
2. your usage pattern. from the output, it looks like you use microb quite a bit and don't consume multimedia / audio content very much, no?

so a thumb vs nonthumb memory comparison would need to be with similar uptime and usage patterns to make sense?

just my 2 cents....

Re: Thumb2 vs non-thumb memory usage comarisions
 

Also i like the general idea to compare 'thumb2 vs. stock', i would suggest to describe the basics (like widgets on the desktops, number of desktops, kernel running, some kind of swap initialization....) to have some ground for the comparison.

Re: Thumb2 vs non-thumb memory usage comarisions
 

I am interested in these results as well.

The test method is very important, though.

Maybe a good and simpler starting point would be a comparison between stock x thumb2 of post-reboot memory footprint of some flagship apps.

For instance, my browserd is reported by
as using 42780. I have not rebooted and I am on standard CSSU testing, with no open windows (I have not checked if browserd is on thumb2, though).

Even better is to use , as it breaks memory usage of the process and its shared libs.

So we could choose maybe 4 standard apps and post their pmap maps of thumb2 and standard, and that would give a good idea of the differences.

Re: Thumb2 vs non-thumb memory usage comarisions
 

Thanks for the edit with the code box Estel!

Re: Thumb2 vs non-thumb memory usage comarisions
 

I think would be most useful if we're comparing the same configuration, one with thumb and one without.

If someone has a spare N900 and a few hours to kill, I think a good test would be:

- complete reflash including eMMC
- disable HAM auto-updates so it does not run apt worker in background while we're trying to take readings
- install CSSU (regular)
- reboot and take memory readings after X minutes uptime (to allow tracker to settle)
- install CSSU-thumb
- reboot and take memory readings after same X minutes uptime

This way we are comparing same to same.

It just so happens I have a spare N900 and maybe I'll try this if I find a few hours and someone else does not beat me to it. :)

Re: Thumb2 vs non-thumb memory usage comarisions
 

Another approach is to compare memory usage of single processes, as this is what thumb2 is about, yep?

We don't care how much of it is swapped or how many widget's one have in desktop - we have list of thumb2 compiled programs available in thumbs repos, and what interest us, is their memory footprint compared to stock (non-thumb) equivalents.

At least, this sound sane for me, or we will drown into hell of non-comparable results. I know that amount of things I've included in data provided via post 1 exceed it - I did it on request, and was kinda sleepy :p

So, unless someone have points I've missed, proper scientific approach would be to take list of thumb2 compiled programs (from thumb2 repo) and compare memory used (absolute numbers), when they're in use, with non-thumb2 ones. This way, no need for 2 N900's per test's contributor.

/Estel

Re: Thumb2 vs non-thumb memory usage comarisions
 

Posting here, sorry i didn't see this thread but have limited time and have had for some time, else i would contribute more.
Also updating the numbers now that i have installed some of the last programs, but still less then before since i did not use some of them anyway....

Normal N900 that is now taking a rest:
Running Power50
RAM: 147MB
SWAP: 107MB
Total: 254MB

Other N900 that is now primary:
Running cssu3 Thumb 14
RAM: 124MB
SWAP: 81MB
Total: 205MB

But all in all i have only install the basics and contacts, g.talk and skype running.
So memory usage is subject to change, will properly post again in a month or something

Re: Thumb2 vs non-thumb memory usage comarisions
 

Another option would be to use something like backupmenu to clone an existing unit without -thumb, and then ONLY update one of them. That would let you (after the update and a couple reboots for safe measure) do a true apples to apples comparison, including not only the kernel, but any updated executable that have been compiled for thumb that may be part of the base install and running. (Like browser and the like?)

That would, IMHO,be a much cleaner and "fair" comparison, and wouldn't require one to lose the active setup on their day-to-day device.

Re: Thumb2 vs non-thumb memory usage comarisions
 

I am thinking about doing this, just need to learn how to use backupmenu.....


On another note, my now primary N900 with thumb holds power like crazy, getting close to 48hours and still over 50% left (59%) this is of course with my dual scud, but other N900 would be just about dead now, even if it has only been lightly used.

Re: Thumb2 vs non-thumb memory usage comarisions
 

I have a second N900 which is almost a clone, that is, it was a clone up to two weeks ago, but on my everyday device some additional programs have been installed like mypaint, OMP and a game or two.

The everyday device runs Thumb2 v.1.4, the spare one that I don't use runs KP50 with CSSU-T v.1.2. If needed both devices can be brought to an almost identical state, by installing the missing apps if needed on the spare one.

In short: If the above is good enough for the OP, I'll run some tests and post the results here.