OpenCL and Skia/Vulkan are turned off, see First steps to take before submitting a bug - The Document Foundation Wiki
It still sounds like an issue with the file but you could click Tools > Options > LibreOffice > View and tick the box Force Skia software rendering, allow LibreOffice to restart and see if there is an improvement in performance.
Guys,
I need to run some chores for a couple of hours. I will be back to this when I get home again.
MS777. Are you a formularisation expert by chance? Know anything about LET functions??
CT
… LET functions are quite new for me, too. But I have ~20 y experience in finding problems in OO documents
Before you submit a bug, it is always good to exclude that the bug is connected to some specific properties of your installation …
Yesterday, while testing ‘=LET()’, something inexplicable happened to me when I tried to enter something. LO automatically converted this function into a Matrix{=…} function block, which has since been missing ‘LET()’.
So far I haven’t been able to get a ‘=LET()’ function to work.
Hey ms777,
Thanks for the offer to take a look at my misbehaving file. Link below. It was slightly too big (4.3mb) to upload here.
I’m not entirely sure how, as I’ve been flailing around with a million restarts and forced quits over the past 24 hours with it, but I’ve managed to get it working through (I think) running it in safe mode. It’s still pretty crotchety though, Scrolls like mollases. “Calculates” (can’t run with AutoCalculate on, heaven forbid) and “Saves” taking a long time.
That’s what I’m trying to ascertain. I am maybe too much of a newb to recognise what I should be looking for, but anyway. it will be very interesting to hear back how well/if it runs on your machine as a comparison test.
A couple of days ago I introduced my first attempt at a LET function. That’s in column BR “Final Exit Event Flag”. That didn’t seem to cause any undue crash tendencies (maybe slow down behaviour, not sure), but yesterday I also introduced another LET function to column AD “Buy Flag Row [in MR]”. I am suspicious that this is around the time when the trouble started.
Thanks again for the offer of help. Much appreciated.
Regards.
CT
Hey Earnest,
It looks very much like:
did the opposite and made matters worse, but I have just read through the Wiki document and ohers linked there too, and I’m on the trail of trying to swap out my current user profile etc.
ms777 above is also hopefully going to take a look at my file as well, so maybe that will zone in on the specific problem here.
If you feel inspired and have the time, feel free to download my file from the post above too.
Cheers.
CT
Hey Koyotak,
Thanks for the additional info.
As I mentioned to ms777 above, I am very suspicious that the LET function may be the cause of my problems. Amongst other things I’m trying, I’m just about to try ridding my sheet of all LET functions now as a test.
I wonder if LET is not quite ready for primetime!!?!
Cheers.
CT
CT-Shares Backtester for ms777-2.ods (167.8 KB)
Hi,
Have a look at this. I have removed the bottom 100.000 lines. I hope I have not removed some lines carrying data. My gut feeling is that the size is the problem. Yor content.xml contained 128 MByte. I have a deeper look tomorrow
Hey ms777,
Thanks for this.
Being a bit of a beginner, I couldn’t work out if the amount of data I was testing with was a small, normal, or astronomical amount of information for the average spreadsheet. This is my first serious attempt at building anything like this, so my inexperience certainly shows.
Great! I’ll take an informed gut feeling at this stage.
Yes, well the final section over on the righthand side (MR Master Reference candle times @ 30 sec) WAS 57K odd rows long, so it’s obviously not a typical “real world” example to cut it down as much as you have done, but it’s an excellent experimental scenario, especially if the amount of data looks excessive to you.
What’s the content.xml size? Is that the raw data? Is 128 Mb a lot, or a little?
I’m just about to embark on a test to put the bloody thing on a diet!!
I’ve been using ChatGPT to help with certain approaches to things I’m less than optimally skilled to handle, and it has mentioned several times that it thinks my Dynamic Named Ranges should be converted away from using OFFSET and COUNTA, as they are “Volatile”. Would that explain the length of time my Calculate (F9)'s take and file Saves take??
Any thoughts on this or other approaches to slimming the damn thing down some??
Thanks again for your help. Much appreciated.
CT
Hi,
I did a little further clean up:
I have removed some 50,000 empty columns on the right side
There was an orphan object on the left side. You can see that in the Navigator pane. I removed that, too
I did not try to understand the logic behind all formula, it is too complex. But here some observations:
You use a mixture of normal functions and array functions (like ={…}, e.g. in col V). I am not sure if array functions are necessary. I recommend to use them only if you are exactly knowing what you are doing.
In your Named Ranges there is a mixture of upper row bounds, e.g. 187, 3617, 97989, 1048576. Maybe this is partly caused by me cutting the rows, but you should look over that. Please note that the handling of complete columns (e.g. B:B) in LO is not possible like in Excel
All office documents are zip files. If you copy your .ods, rename it to .zip, open, you will find the content.xml I was mentioning
Good luck,
ms777
CT-Shares Backtester for ms777-3.ods (182.7 KB)
P.S.: There is no indication that the LET function is the culprit …
As of today, 2025-04-20, that is possible. It was implemented several versions ago.
Ýou are right, thanks for pointing that out. I completely missed that …
@Clearwell_T: That will probably simplify some of your Named Ranges / Formula …
LIke the vast majority of today’s spreadsheet users, you try to implement a database on spreadsheets.
Hi ms777,
Thanks once again for the help.
Aahh, yes, that’s one thing that intrigued me ie: why the default document has so many columns (see my answer to Villeroy below). I had meant to get around to removing them. They weren’t my doing.
Actually, that was a place holder for my lame attempts at a horizontal bar chart, to display trade times or some such. Thank you very little for being so callous about my carefully constructed artistic piffle!?! 
The only formulae that are using array enter, are precisely the ones that need it. I’ve been very meticulous about testing whether array enter is necessary. It seems to me that it’s very obvious when it is and when it isn’t required.
That’s definitely mostly a result of your cutting. Although I was trying to aim for a sensible number on the DNR spans for each section. Small, medium and large row number requirements. I am just in the middle of trimming down the DNRs to the bare minimum, and rationalising the lengths again in light of further experience and your helpful comments. Ta for pointing this out.
I’ve seen the update to this below.
Great, but I’m still unsure what a “sensible” size for this file would be??
Yes, I think I was starting to come to the same conclusion. I replaced my User folder, and I seem to be making better headway now, modulo trying to handle too much raw data. See my comments below.
Thanks again ms777. Appreciated.
CT
Hey @Villeroy,
Thanks very much for the reply.
I’m really intrigued by this comment. I’m totally sure I am one of those people, but only through shear self-admitted ignorance.
LIke the vast majority of today’s spreadsheet users …
This seems like the most telling part of your statement. Does this mean that there isn’t a suitable product that’s “fit for purpose” out there in the market, or just that, like me, inexperienced users such as myself, assume unattainable results by focusing on the wrong product solution??
try to implement a database on spreadsheets.
What I find most misleading in this scenario as I find it is, why would the default configuration of a spreadsheet include over a million rows and north of 50,000 columns, if that wasn’t intended to lead neophytes such as myself into believing that there was some serious raw capabilty under the hood just itching to be unleashed.
My failing no doubt.
Can you recommend a suitable equivalent database version of what LO does, that I should look into?
Thanks again for your help. Much appreciated.
Regards.
CT
Did you save the intermediate copies as new files: “safe as…”? Then find the most recent file that works well. If you only have one intermediate file that you’ve always overwritten, then you’ve made a typical user error. If you’re extremely lucky, you might find the cell with the programming error. Or you can go through the program code line by line, counting the parentheses, and spot a typo or missing part of a line. There are programs in C++ that let you test your code for errors. I doubt LO programmers have developed anything similar. Therefore, I recommend you will copy the contents of the table and paste it into the new blanc table as plain (unformatted) text to at least save the sensitive data. A very nice job for rainy holidays 
Hey Koyotak,
Yes, although I’m a total LibreOffice newb, I’m long in the tooth with file management experience. I have a long sequential list of Save As…'s. The problem so far has been actually working my way back through them to find which will open, whilst simultaneously troubleshooting from other angles with the help of others here.
It very much looks like now that I’ve managed to get the latest “Save” open, I assume they all will open, so not much to be gathered there, unless the issue DID creep in at a particualr point in the day.
As discussed above, I’m going on the hunt for eradicating LET functions as my next port of call.
Thanks again for joining in. Much appreciated.
CT
Hallo @Clearwell_T
I used to be a maintenance technician, but unfortunately, that’s no longer valued these days: “If only you could have found the error faster, earlier, and more easily!”
I circumvented a non-obvious error by splitting the suspect entity into three parts and testing each one individually. If the error occurs in all three parts, I move out and on to the system. Before I poke around in the system, I question as many people as possible who have handled the system, including events that deliberately don’t fit together with something else. Putting two things together that don’t fit together at all is my special talent, one that only a few people possess.
Hey Koyotak,
I’m a recovering recording studio engineer, and it takes a certain breed of person to excel at that job (not saying I did) too.
That perfect blend of technical competence, musical understanding and a psychology PHD is needed to avoid many “customer exempt (of course 
)” session derailing problems.
Thanks again for helping keep me in line.
CT
High @Clearwell_T
I welcome any engineer/manager who stays in the background, is willing to experiment, and wants to see the broadest possible picture, unless we cooperate well. Well, I’ve had very few of that kind, and I’ve learned nothing.
Except that that good symbiosis was a catapult that shot into progress, 10 years into the future, as an architect once explained to me. Since then, I’ve been trying again and again to shoot myself down and bring myself down to my current, also stagnant, position.
I’ve had many of the other kind, the ones with arrogance and terrible patronizing behavior, even those with narcissistic tendencies. The collaboration was disastrous, but I learned a great deal from them. I mean that explicitly in terms of character and human qualities, the effects of fractures and injuries. A significant advantage in old age
I hope and wish that you were at least able to save the valuable data, text, information, and values. The layout and formatting with the recessive calculations, however, are much easier to regenerate and can be better structured the second time around.
Have a good time!