

Sorry, my English is not good enough.
ZSWAP basically crashed. The computer had no stutters, it stopped working at all. It did not really feel slower with ZSWAP vs. ZRAM.
I used the same computer with exactly the same desktop/browser configuration, so it should be a ‘fair’ comparison between ZSWAP/ZRAM.
I also have to add, I am using ZRAM on high memory devices, too, w/o any trouble. ZRAM should IMHO even perform better, when you have more memory, because the RAM disk can be bigger and hold more data.
Sorry, I cannot follow you.
There is no reason at all, to not have physical swap when using ZRAM. My physical swap just has a lower priority and is never used, when using ZRAM.
You claim that there is no performance difference, once you have >= 8 GB of RAM. I would assume it is much relevant what kind of usage is thrown at ZSWAP/ZRAM instead of the size. Perhaps you could elaborate, why, in your opinion one will not observe a difference in performance and why this limit would be exactly 8GB? I also use ZRAM on my 8GB machines, with a physical swap partition as backup and physical swap is never touched.
OOM killer is IMHO independent of ZSWAP/ZRAM, because swap space has an upper limit anyway, and a functioning OOM killer is IMHO preferable to a crashed system. I was just surprised, that the OOM killer did not work for me, when using ZSWAP, because IMHO it should have killed the browser in my example.
Finally, why do you claim that ZRAM performs better on low-memory devices, because I do not see a reason that it should, or that such a generic answer makes sense.
Please understand my intention here: I want to understand WHY you make these statements and what is the technical reasoning behind them.