I have a couple of spreadsheets open during the week. CPU background usage when I just open LibreCalc is about 0.7% (when LibreCalc not in focus and not used). Then the background CPU usage always goes up to about 5% when I’m doing nothing on the spreadsheets and they are in the background, not on screen. The spreadsheets don’t have any calculations going on, no macros. What is LibreCalc doing with that high CPU when it should be idle? How do you find out?
It would be interesting to know your OS and your version of LibreOffice. Copy and paste this information from the Help > About LibreOffice window.
For me, WIndows 11 and v7.6.4.1, the CPU usage is 0.1% and it does not matter if Calc is the active window or in the background or minimized.
Version: 7.6.3.1 (x86_64)
Env’t: CPU threads: 8; OS: Linux 6.4
UI: UI render: default; VCL: gtk3
Locale: en-US (C.TUF-8); UI: en-US
Misc: Calc: threaded
Again, the higher CPU usage shows up after I work on the spreadsheets and after they are open after some hours. I haven’t clocked it, but I always see 4-5% CPU usage after a while.
Correction, actually. LibreCalc is always in the foreground of the systemd-nspawn container that it’s running in, it’s the container window gets minimized and out of focus (with LibreCalc in focus inside of it).
So, I just relaunched LibreCalc sheets and will try to keep track of CPU time. Right now, soffice.bin is consuming 0.0% CPU when the window is not in focus, and CPU usage goes up when I focus the window. Normal expected behavior for now, let’s see for how long.
Well, it’s been 4 hours since I launched Librecalc, and now it’s starting to give me 4-5% cpu utilization. It may coincide with me launching a python script that loads an .xlsx file that’s open in Librecalc, but when I stop that python script, the utilization remains high at 4-5%.
High background CPU utilization does not stop as I sequentially close the files that are open. It stops only when the last open Librecalc file is closed and all that remains is the Libreoffice splash screen. When I reopen the file(s), CPU util is at 0% again. So it has to be tied to the files that are open for a while somehow, which doesn’t make sense.
It may make sense if AutoRecovery is enabled with a rather long interval between saves.
Byt the way, considering the huge diversity of Linux distros, what is yours? And which desktop manager?
Fedora. Nspawn container. And just a plain openbox window manager. Really no issues before, solid performance, never crashes, just noticed this weird background CPU utilization uptick. Autorecovery is off.
Please, be accurate Fedora xx? And don’t open multiple posts just for a few lines. You can edit previous posts. Be friendly with potential contributors. They’ll appreciate not being forced to scroll long just to gather small bits of information. Make your posts concise and “grouped”.
Fedora 39.
I disagree with you, editing former posts is not friendly with actual contributors and new contributors will not understand why some questions are asked while the answers have already been mentioned in previous posts.
This site is not a forum. It is a Question & Answers one. There is nothing like a chronological conversation. The underlying Discourse engine can reorder contributions according to its own relevance criteria. The consequence is worse than you think: since chronological order is lost, the “thread” becomes completely un-understandable.
The idea in the site is to expose a detailed problem at head of a topic with all needed information. This helps others to evaluate if your problem is theirs, without scrolling and fumbling through numerous additions.
Answers provide several ways to solve the issue. They must be concise too so that visitors don’t lose time to explore how the solution built up.
When a topic grows beyond some threshold, the Discourse engine offers a button to summarise the whole topic so that the concision goal can be met. Comments are not meant to be kept once OP and contributors agree on a solution. In theory (but this is rarely done), a solved topic should be cleaned to keep only the question and its answers. Comments should not be necessary to understand both the question and the answers.
1 Like
Well, having answers written before questions via edition of posts does not seem natural for me. And there is a chronological concept, look at the time to the right of the header.
Here is a summary provided by Discourse: as you can read not very smart summary!!!
Okay, I may have figured this one out. This needs to be fixed!
So, basically, when one of the open librecalc documents has anything copied into clipboard, I get this annoying ~5% (fluctuates) CPU background usage. The document with the copied text does not even need to be in focus, it can be in the background, it still produces the CPU usage. I think it has to do with drawing that live border around the copied selection, and the drawing takes CPU cycles. When I ESC and the copied selection is reset, the CPU usage is back down to near 0.
Now, I am running librecalc in an nspawn container and in Xephyr, so this may be related to CPU being used to draw things, but either way there’s something nasty about drawing things in the background…and even in the foreground…with that much CPU. Maybe decrease fps on copied selection border too?
It is as long as you use comments to the same main entry - be it a question’s comments, or an answer’s comments. But the different answers themselves will get reordered based on their “value”, and their order will be not chronological. In general, it’s best to consider the advises given by high-profile contributors, they are usually given for a reason; yes, we realize that you are a smart person, but assume that the site is created with some structure / goal in mind, and that the advises could help the wanted structure. Note e.g. how your “solution” got hidden in this comment thread (because, well, comments are something auxiliary, not as important as answers).
Reading the site’s rules may help: there is a proper place to file bugs. However, you may see not something related to LibreOffice, but maybe to some clipboard manager, which might request the copied data in a graphical format, so LibreOffice has to generate the image of it for every format it gets requested.