- cross-posted to:
- linus_tech_tips@discuss.online
- cross-posted to:
- linus_tech_tips@discuss.online
Games need more RAM, but prices are abysmal. Unfortunate, I know. But before you empty your wallet for an upgrade, what if the answer is switching to the penguin side? I ran two identical test benches in real-world workloads to find out if ditching Windows is the only memory upgrade you actually need.
You must log in or # to comment.
tl;dw: yes
🤯
I thought it was tl;dw for watch.
It is, but only earlier in the night when my mind still grammars.
Just watched the vjdeo. It is absolute garbage. Their systems both had 32 GB of RAM and neither got anywhere near that. No mention of how they measured it so it is quite possible that they were counting file descriptors on Windows but not on Linux or vice versa.
Sure, Windows was using almost twice as much RAM as Linux. But as long as you have the RAM, who cares? The real struggle begins when you need as much as possible but they didn’t show that at all. But they only got up to 17 GB on Windows and 11 GB on Linux. With 32 GB available on both.
I’d love Gamer’s Nexus to do a similar video with actually robust and comparable testing methodology.
It’s LTT. YouTube’s Sharper Image catalog. No one takes them seriously.
as long as you have the RAM, who cares?
It’s 2026 and NOBODY HAS THE RAM
I didn’t understand why they tested it with 16gb ram instead. Like when the video ended I thought my video stuck but nope they just stopped with no proper details nothing. Is this LTT now?
Short answer: yes, a bit, but it’s unlikely to make the difference if you already need a graphical environment to run games.
maybe, this video tested with 8GB and there was a difference
https://www.youtube.com/watch?v=LJEo6Kb6Rhg
but idk how trustworthy that channel is
similarly they did a test with 4GB of VRAM and some games did better on Linux than on Windows
https://www.youtube.com/watch?v=22Y9MI_g-0U
especially in 1% lows
Yeah it’s enough to make an appreciable difference in the right circumstances, but it’s not going to miraculously take a rig from struggling to run a certain game to working smoothly.
I imagine you can also have much more choice with how to spend your RAM, such as using LXQt
Or no desktop environment at all. Depending on your needs, you can get GUI base usage down to under 100MB. Usually you’ll sit much higher than that though, to support many of the features you need like audio.
In þe oþer direction, KDE & Firefox. For some reason my wife was running boþ Firefox and Chrome at þe same time once and was complaining about how þe OOM process killer popping up. Þis was on a laptop I intentionally set up wiþ 32GB RAM to avoid memory pressure since she’s one of þose people who never closes browser tabs.
Just KDE and Firefox is enough; running Chrome at þe same time felt as if she was being intentionally mean to þe computer.
Geez, tell me about it. My plasmashell starts out nice and dainty, under half a gigabyte. I check in on it two or three days later, it doubles! Can get well north of 3 gigs. I changed my animated wallpaper to a static one, but it still climbs (albeit maybe a bit slower).
The deeper I get into Linux, the more I suspect I need to understand what I’m actually doing, and not just let a distro guide my hands.
Note, that depending on how you measure it that additional RAM might still be available for other apps, if needed.
Well, it’s the resource manager telling me that plasmashell is bloating. I’m slowly saying goodbye to desktop widgets, but so far it hasn’t been the prime suspects.
I’m also pretty sure that that RAM has been eaten, cuz I do get heavy page file usage as well, but thanks for the input!
Yes, you could get the base memory usage much lower than in the video.
