what are the downsides of increasing swap size

I am working with very large datasets. In windows I was able to load them into RAM and it would use swap memory, which would be 5 times faster than using the fastest disk based dataset storage like deep lake, and I would even be able to continue watching youtube while it is running. In bazzite I notice I can't even reach 1/3 of the dataset size before it becomes impossibly laggy. And this is likely to swap size being 4 GB, where in windows I am pretty sure it was set to very large value like 64 GB by default. So I googled "bazzite increase swap size" and stumbled upon the following thread: https://www.reddit.com/r/Bazzite/comments/1dlbf00/increase_swap_memory/ The answer states that "I typically wouldn't recommend setting it above half of your total RAM ". And I was wondering why that is. Because I would like to set it to the same value as it was in windows.
1 Reply
1
1OP2mo ago
I changed it to 32GB and it worked, except I would like to know what are the downsides because maybe I will change back to 4 GB after I am finished with my datasets somehow it still uses 2 times more memory though like it already filled my 32GB swap okay then I set it to 64

Did you find this page helpful?