Sometimes, my computer is very slow and sometimes really fast. Fedora 43.
Hi everyone!
My daily driver is a Surface Go 1 running Fedora with 8GB of ram and 128GB of storage. It is always hooked up to a Philips 273B screen via USB-C.
Most of the time, it is really fast and a perfect tiny Linux and Gnome machine easily hookable to a big screen for when you're not travelling.
However, sometimes, after installing updates but maybe not always, it is slow as hell. Sometimes, detaching the Surface from the big screen and hooking it again, solves the issue, but not always. It is a behavior I already had when I was using Ubuntu and I've had on Fedora since version 36.
Here are some useful printscreens from HTOP and the ressource management system:
#high cpu usage
#low cpu usage
I thought that maybe installing the Surface kernel would stop the issue, but it didn't..
Sometimes it's annoying enough to make me just want to use my wife's MacBook Pro 2012 running Fedora as a daily driver but the form factor is less practical.
Thanks in advance for any help!
Edit: it happens on startup, even after days of inactivity
Nanook
in reply to Dariusmiles2123 • •jokro
in reply to Dariusmiles2123 • • •IanTwenty
in reply to jokro • • •This is my thought too. Your screenshots don't show cpu freq. Everything can appear normal on other measures but if your cpu has throttled right back it'll run slow with no obvious cause. I have seen this on certain laptops when external peripherals are added/removed after waking from sleep. Sometimes forcing sleep/wake or a reboot will fix.
Use
lscpuat the cmdline or better yet install gnome extension 'system monitor next' and put the cpu freq graph in your top bar to watch it in realtime.extensions.gnome.org/extension…
system-monitor-next - GNOME Shell Extensions
extensions.gnome.orgsnailstone
in reply to Dariusmiles2123 • • •kumi
in reply to Dariusmiles2123 • • •Apart from what others said about power/throttling, I wonder if the filled up memory during the upgrade (or other memory-heavy use) pushes some central pages to swap and then they stay there after?
After the upgrade and you have plenty of free memory again you can force back everything to RAM by temporarily disabling swap:
To list swap devices, just run
swapon.Also switching to an X11 window manager can be quite a lot snappier than modern GNOME for older hardware. You could try Xfce, Cinnamon, MATE, or KDE with the X session.
If it's not throttling/thernals, I wouldn't be surprised if those two together is what made things worse after migrating dist.
If you've been swapping heavily over time you might also want to check disk health with
smartctland check that you don't have related errors indmesg.If you press tab in htop you can also see if there is high IO load going on.
atzanteol
in reply to kumi • • •kumi
in reply to atzanteol • • •Hence:
If it's like the last htop image should be no problem.
atzanteol
in reply to kumi • • •Turning off swap could make things much worse though. The system will have less memory for file caches.
I'd leave swap alone, just monitor for whether the system is paging frequently. "vmstat 3" should show if you're writing to swap frequently.
kumi
in reply to atzanteol • • •How so, given that we immediately re-enable the same swap device right after so that it's only off for a very brief moment? Let go 😀
Anecdotally, this maneuver can help tremendously tonrecover responsiveness in some cases. I guess the overall sitiation could be improved by tweaking
vm.swappiness.Nanook
in reply to kumi • •atzanteol
in reply to kumi • • •See that
buff/cachecolumn? That's memory being used by the system for caching. Files you you open and access get cached into memory as do inodes, filesystem objects, etc. If you run a "find / -type f" twice in a row the second one will be significantly faster because the first run cached a lot of objects into memory.By starving the system of memory all that will be flushed and you get more disk access doing things you're actually trying to do. Whereas things sitting in swap are there because they aren't currently needed.
By turning off swap and then back on again you're just forcing the system to drop all that cache which it will then attempt to reclaim space for and push things back out to swap.
I don't know what benefit you think you'll gain in the process.
mech
in reply to Dariusmiles2123 • • •I'd guess at overheating, or an error with the CPU's energy management.
First thing I'd do is look if there's a BIOS update available.
Then install lm-sensors (there's also a GUI frontend called psensor and a gnome shell extension) to show CPU temperature, and check if they're too high. If they aren't, you could set your CPU to always use max power and see if that fixes it, but it will reduce battery life.
Or try a different distro from a live USB and see if that makes a difference.
isgleas
in reply to Dariusmiles2123 • • •utopiah
in reply to Dariusmiles2123 • • •It's all just speculations, both what you suggested and what others said.
You are on the right path with your screenshots but you might not be measuring the right thing.
So, you need a (paper) notebook to record objectively (not your biased feeling assuming a pattern that might not exist) when it happens and for how long. Only from then can you backtrack to WHAT causes it. Sure you can have some hypothesis (update related, screen attach/detach, BIOS, RAM, etc) but that should NOT lead to your data acquisition.
So you htop is nice but AFAICT it's just about CPU and memory, it's not about e.g. IO so consider instead
iotop, in particular if one process is some indexing (e.g. locatedb). Theoretically if it's not CPU/memory (which you are saying it's not the case) then it basically just leaves IO, that can be again indexing, some heavy process that is bottlenecked on disk access, but can also be a bug, e.g. BT pairing/unpairing that happens faster than you can notice.Think of this as a fun investigation that leads you to better understanding of your setup, good luck.
Nanook
in reply to Dariusmiles2123 • •