Linux Swap Partition Usage?

I think this is in the right forum (I really don’t know where else it would go…), but here it goes.

Windows and Linux both have their virtual memory (Page File on Windows and Swap Partition on Linux). As far as I can tell, windows specifically reserves it’s paging file for whole idling applications. Windows won’t dedicate an application to the page file unless it is idling and possibly minimised as far as I know. (I believe this because of my tests with higher memory usage baking with blender)

My main question here is, when selecting a resolution for baking a fluid simulation in blender, it gives you an estimated memory usage for that resolution. Baking at a resolution of 512 gives a basic fluid simulation an estimated memory usage of around 14GB. The average home user or novice 3d modeler most likely isn’t using an extremely beefy system to play around in blender, so my question is, really how does Linux use the swap partition?

Does linux use the swap partition like windows, or would creating a swap partition of like 40GB allow me to bake these extremely high resolution fluid simulations without crashing blender from lack of memory? Is purchasing a whole ton of ram (heck, a new server hardware system) the only way to bake these extremely high resolution fluid simulations?

Windows and Linux basically do the same, and that’s called paging.
Neither do actually “swap” whole processes anymore. Your whole physical memory is divided into pages (4KB usually), and whenever more memory is needed, some of those pages have to be swapped.
Strategy may differ a bit on Windows and Linux, but it will use criteria like last use of the page, the frequency of use etc.

Though with a 32bit system you will definitely NEVER get more than 4GB of memory per process, since you simply can’t address it. Actually, the program will get killed somewhere below 3GB on Linux, and even earlier on Windows.
With 64bit Linux you theoretically could put those 40GB in swap, but prepare to have up to 1000 times longer computation time, because that’s about the difference in access time between RAM and HD…