Honestly, Linux distros ( CatchyOS in my case) are getting so good at managing ram that now i’m just curious to know if i could use that extra free ram to remove some work load from my old pc cpu.
If you have spare ram, consider setting up and using a ramdisk. A ramdisk is just what it sounds like: a filesystem made of ram. You can store anything on it whether it be something small and temporary or large and less temporary. You can also use it to enhance the startup of things, but ssds are so fast these days anyway :)
As others noted, the only way to improve CPU performance is either by making the CPU/RAM actually faster (higher clock speed, better/newer CPU, faster RAM so the CPU wastes less time waiting for the data read) or by wasting less of the CPU resources on useless/needless work. Caching helps in the later case, like waiting less time for reads that are cached in memory, or avoiding recomputing something you already computed before. However, Linux already has the page cache that uses free RAM to cache disk reads, and using more memory to cache computation results is a compromise that has to be decided by the software developers before it’s even shipped to you, so not much you can do there.
No, those parts do completely separate things.
I really think someone else needs that ram more than you.
If I was less lazy, I’d make one of those “I consent, I consent, isn’t there someone you forgot to ask” memes with a guy putting a stick of RAM into their computer and the guy and the computer both say “I consent” and Sam Altman in the background says “I don’t”.
I did not make this, found it half a month ago.

To be fair, it’s entirely possible I had run across the same meme and it was stuck somewhere in my subconscious and then I thought I’d come up with it when in fact I was just remembering it.
download more ram
In what way were you thinking? RAM can’t execute instructions and stuff like that, you can use spare RAM capacity to cache files and stuff so the cpu doesn’t have to wait for it to be read from disk, but linux already does that, what else is there?
You could rent it out to Sam Altman, inflate the market value of RAM, revalue your company according to the new prices, sleep with a chinese spy, tank your stock prices, get the company “under water” and cause the ai bubble to burst. In that order.
CatchyOS
It’s really weird that so many people are hallucinating a t in this name. It’s Cachy. Like cache.
It’s not very catchy.
Look it’s dogs fault I’m dyslexic
Same 👍
Yes. Merely by having it free, that means more data can be cached in ram and not have to be loaded from elsewhere, which takes (a very small amount of) CPU time.
Think of it like this: you’re a student in a high school trigonometry class, and you’re taking a test. You have a scratch pad for jotting stuff down temporarily while you work, you have a test sheet to put your answers on, and you have a calculator for doing sines, cosines, and tangents. Could you use the scratch pad to lessen the workload for the calculator?
To some extent, yes; you can break down the calculations into smaller workloads and write them down on your scratch pad, then type them into the calculator one-by-one and put them back together at the end. You could also maybe remember the answer to a problem you’ve already solved earlier and write it down here. But no matter how big and nice your scratch pad is, no matter how well you write the problems down on it, the scratch pad will never solve those problems for you.
In computing, the test-taker is the operating system, the screen is the test paper, the calculator is the CPU, and the scratch pad is the RAM. So you could break down the calculations into smaller workloads and put them into RAM, then run them piece-by-piece and put them back together at the end; but every OS already does that. You could remember the answer to previous calculations and output it from RAM; but that’s called caching, which is something else that every OS already does (and something that CachyOS in particular focuses on).
These are on by default already; the only way you can really help it is by minimizing open applications, which increases the amount of RAM that the CPU has available to write to.
But, no matter how much RAM you have, no matter how well you write the workload to it, the RAM is just a scratch pad. It can’t process anything for you.
Maybe disabling zram.
Pre-calculation or calculation caching is about the best save you’re going to get, but each and every piece of software would need to be rewritten with that idea in mind - assuming it wasn’t already written that way in the first place - rather than it being something the system can impose.
It’s plausible that someone could write or modify an OS that specifies to each program what they’re supposed to do, be that optimise for time, optimise for CPU, optimise for RAM, etc. etc., but the program itself still needs to provide the ability to do each and every one of those, which is time and effort that program writers may not be willing to put in.
For a more concrete example, although still something of a toy, there’s a program in the
bsdgamespackage calledprimesthat’s ~80k in size.Clearly, at that size, it does not contain all prime numbers but it sure can generate more than 80k of them quickly. It is highly optimised code, if my efforts at beating it are anything to go by.
If you, as a hypothetical prime number aficionado, wanted even less work on your CPU, you could store your database of prime numbers in RAM (or disk) and not generate them over and over with
primes.primesdoes not do this. It is not written to do this. It could be, but then it would be a larger program to account for extra modes of operation, which is more work for the maintainers.Yes, though usually it’s harder to make that tradeoff than it is to avoid disk I/O contention or network bandwidth.
RAM-based caching of disk contents will have some minimal impact that way, though the main goal is reducing disk I/O; using something in RAM will use slightly less CPU time than pulling it off a disk, especially if you are using an encrypted filesystem or encrypted block device. Your system will already be using otherwise-unused RAM to cache content on disks.
If application software is smart enough to have its own caching based on free memory or based on a number that the user uses, you can go twiddling that. IIRC, Bethesda’s engine (Skyrim, etc) has some setting that lets one keep a given number of cells loaded (though I don’t know whether all loaded cells are active or not, so it might be aimed more at limiting how many areas the computer is dealing with). But that’s at an application level, not a systemwide toggle.
There isn’t really a great systemwide Linux feature that I’m aware of that gives you major CPU usage reduction by expending memory. The problem is, I think, that in general, if you want to avoid costly recomputation, it’s often preferable from an application standpoint to use a disk cache rather than a memory cache. If the cost is large enough, it’s preferable to just use a disk-based cache, which permits for persistent caching (I don’t wipe my cache every time I terminate, as a process) and usually permits for larger cache and still gets some of the benefits of RAM-based cache by letting the kernel handle sticking frequently-used disk contents into RAM. Plus, SSD storage has far lower latency (and somewhat-higher throughput) than rotational disk storage, so the cost of using a disk cache versus a memory cache is lower than it once was.
Firefox is one program that I can think of that has both a disk-based cache and a memory-based cache, and looking at my current
about:cacheoutput on a new Linux laptop, it looks like these days, the default is to have the memory cache off by default, as I’ve got no entries in the memory cache.





