Does on mean that /sys/kernel/mm/lr_gen/enabled returns 0x7?

Just set vm.page-cluster to 0, to disable the page read-ahead (AFAIK the default value was set when swap was still on rotating disks).

Arch wiki has some good options for tuning the VM

Debian does not enable it by default, cat /sys/kernel/mm/lru_gen/enabled will be a 0.

I am running Debian 12 on all of my devices with Debians vanilla kernel! :-) Just enable MGLRU on Debian like it is described in this blogpost.

One further tip for ZRAM: On my device the LZ4 algorithm was noticeable faster than ZSTD (didn’t try ZSTD with the enabled MGLRU, yet) and it was important to disable the RAM page read-ahead on my device.

I don’t know either, but unless one uses zstd (lzo seems more like a thing for this hardware), I would hope that it is totally usable. (Running zstd memory compression on a Raspberry Pi 2, w/o any noticeable speed impact)

Show me some numbers! ;-) … Perhaps I miss something, but basically we have 32bit pointers vs. 64bit pointers, the rest of the data should be the same size. 64bit should be faster for tasks where the CPU is the bottleneck/computations, so IMHO it will be an interesting tradeoff with no clear winner for me.

The most important thing is not the distribution, but to enable ZRAM (or ZSWAP) and use a lightweight desktop. I am not sure how much difference a 32bit vs a 64bit distribution makes, but if possible you could take one for the team and run some trials and report your numbers (RAM usage) back here.

Of course I recommend Debian with a lightweight desktop of your choice, or Alpine.

Thanks, this would indeed solve my problem. Still hoping for a better solution, but if everything else fails I’ll utilize it!

Thanks, that would be a valid approach and my last resort.

As you said, I hope someone knows a more elegant solution, though!

Who else can survive for years on eating their own foot-skin? :-P

Great pick!

So many quotes from Casablanca are part of our everyday culture… and AFAIK it is the first movie ever to feature a flashback in a flashback. Combined with the awesome cast of actors this is a masterpiece. :-)

One is not enough, and a lot of great movies where already named, still, some great movies are missing:

• Heat (Michael Mann) Every single time I see it, it is brilliant and I discover something new
• Jin Roh (The original animation movie), awesome atmosphere and only after the 2nd viewing one can really appreciate it
• Near Dark (1987) Why the hell did nobody ever produce something like this ever again?
• Miami Vice (Michael Mann), ‘Style over substance’, in a great way, although I have the shaky camera
• Seven Samurai
• Casablanca
• Strange Days
• Point Break
• XXX (Nobody understood that it was a parody back in the days :-P)
• What we do in the shadows
• Brazil
• Rocky
• Eternal Sunshine …
• The city of lost children
• Leon the professional
• Dolls
• The Killer (The original of course)
• The last unicorn
• Dark City
• The thing
• The Lost Boys
• Spirited Away
• Donnie Darko
• Rashomon
• Brother (2000)
• Parasite
• Hatsukoi (First Love)

… from the top of my mind. :-P

Nice, thanks a lot, especially the dirty_bytes settings are interesting to me, because I experience hangs with too much disk IO :-P.

Cheers!

Could you ELI5 the last five settings? I saw that Chrome OS sets vm.overcommit_memory = 1, it seems to make sense but is missing here.

Thanks a lot! You are right, I saw this already.

I can confirm the findings with my benchmarks: zstd has the best compression, lz4 is the fastest.

FreeCiv is a classic and still fun. rogue and nethack are good, too.

To my understand it is swap read-ahead, and the number is a power for the base 2. This means the default reads 2^3 = 8 pages ahead. According to what I read, the default of 3 was set in the age of rotating discs and never adapted for RAM swap devices.

Thank you for your answer and your insights.

In my unscientific tests, sysctl/vm.page-cluster made a measurable difference (15% faster when setting it to 0), and it seems everyone else (PopOS, ChromeOS) tweaks at least this setting with ZRAM. I would assume the engineers at PopOS/ChromeOS also did some benchmarks before using this settings.

Now I really would be interested, if you would measure a difference on your 1gb potato SBCs, because IMHO it should even have a bigger impact for them. (Of course, your workload/use cases might make any difference irrelevant, and of course potato SBCs have other bottlenecks like WiFi/IO, which might make this totally irrelevant.

there is no android phone that I am passionate about,

Not what you asked for:

A phone is a tool which should enable you to do stuff. Be passionate about friends, hobbies, art, not a piece of plastic.

Being forced to use iOS (work phone) and Android (Samsung, also work), both suck IMHO but Android sucks less.

My next Android will be a Pixel, as others suggested custom roms are the way to go, but even vanilla Android is more functional/open/practical for my needs than iOS.

I would never buy Apples shit with my own money: Dumped down, locked down and in the end you are renting a device from Apple to pay fees for their Appstore and Cloud offerings and vendor lock in. No thanks.

• Another vote for “Slay the Spire”, have bought it for all platforms and wasted too much of my time
• Horizon Chase Turbo is a very good arcade racer
• Danmaku Unlimited 2 and Danmaku Unlimited 3 for bullet hell shmups

Not arcade style, but Final Fantasy I-VI pixel remasters are IMHO great on phones.