You’ve probably heard — we’re currently experiencing very high RAM prices due mostly to increased demand from AI data centers.
If you’ve been priced out of new RAM you are going to want to get as much value out of the RAM you already have as possible, and that’s where today’s hack comes in: if you’re on a Debian system read about ZRam for how to install and configure zram-tools to enable and manage the Linux kernel facilities that enable compressed RAM by integrating with the swap-enabled virtual memory system. We’ve seen it done with the Raspberry Pi, and the concept is the same.
Ubuntu users should check out systemd-zram-generator instead, and be aware that zram might already be installed and configured by default on your Ubuntu Desktop system.
If you’re interested in the history of in-kernel memory compression LWN.net has an old article covering the technology as it was gestating back in 2013: In-kernel Memory Compression. For those trying to get a grip on what has happened with RAM prices in recent history, a good place to track memory prices is memory.net and if you swing by you can see that a lot of RAM has gone up as much as four times in the last three or four months.
If you have any tips or hacks for memory compression on other platforms we would love to hear from you in the comments section!
Reminds me of RAM Doubler on System 7 or SoftRAM/MagnaRAM on Windows 3.1! 😂
Linux world these days really is 90s all over again! 🤣
If only. Bring back the 90s! :)
Oh God no, it was horrible. Net neutral at best. Tried, tested, trashed. Everyone hoped it was going to be like SMARTDRV except for your RAM, but…no.
Windows 10/11 does exactly the same, it even shows it in taskmgr.
Is not better to remove the sh*t from the kernel? is not normale that the kernel use all this space… Then you can optimize it with memory compression, but what the point to have a lot of unused stuff in the kernel?
Like what? You’ve heard of kernel modules, right? What “unused” stuff are you thinking of?
I have zram enabled on all my systems, so much unused CPU power it just makes sense…
Especially on embedded systems, with overlays to reduce/eliminate flash writes
Just a heads up for Pi users with Pi OS, In trixie a recent update has added rpi-swap.
this conflicts with zram-tools. Zram-tools service fails to load, I found out the hard way
Why would you do that when you can simply visit downloadmoreram.com
Don’t even use the RAM I have. No need for compression. The only thing the RAM prices have done is stop any upgrades.
for real!
This article reminds me of a similar hack i used back in like 1996. Used an ‘after-market’ patch to Linux 1.2.13. I “needed” the compression because i had 4MB of RAM and it simply wasn’t enough. A year later, a summer internship paid for 32MB of RAM and i never worried about RAM ever again.
Anecdote about the kernel hack…back then, the kernel stack was one 4kB page and the compression code would sometimes blow that, eating the 0xdeadbeef magic bytes at the end of the stack and causing an Oops. That really amused me, but it was an easy fix to just use two pages for the stack. I was sure glad to get rid of the hack though.
You’d think with ssd using page files and disk cache
Itd be actually more useful than even with a 10k rpm hdd
1TB ssd?
64GB of page memory…why not?
I mean ssd speeds can be similar to some ram
Nvme and m.2 now instead of just SATA
And at least you can shutdown the computer and hibernation uses zero power
Since it just restores ramdump and registers from ssd
Not even close. DDR5 is around 40GB/s or even more. And the latency difference is an order of magnitude slower.
Well wit the kinda of money ai got to inflate dram
Just use sram, nobody going to complain about sram prices, if anything they go down
Can’t get faster than sram that can run at CPU clock
Back of the napkin, MRAM would be interesting, and available. Ships from Taiwan.
They can afford the several grand per GB cost…
And asynchronous and no memory refresh cycles to waste clock cycle
Just spit an address get some data, limiting factor is propagation delays
How much 86_64 runs without wait states for memory…
RAM price squeeze is a nice way of putting anticonsumer monopolistic megacorporate market fixing. What the AI boom is doing to the PC market is criminal. Huang is a pos
Heh, I seem to remember practically everyone complaining about how fat and inefficient programs were. Now is the time to put that complaint into some action and put programs on a diet.
I’m not using ram definitely
But you used some RAM to post this comment though
For me it’s WOM…
Every comment they’ll ever make, stored in ROM. Nothing new under the sun.
It’s called swap. It’s a good thing. Swap to SSD works very well, since the latency is often 1000 times lower than for HDs.
Sure, thrashing is bad, but zram doesn’t help that at all – even makes it worse, since compression and decompression is more expensive cycle and joules-wise.