Something else was running in the background during your disabled test because it makes zero sense that using a feature that creates more reads and writes to the nand flash would actually increase a nand flash read and write benchmark performance. Logic dictates that your test is a fluke/invalid.
Something else was running in the background during your disabled test because it makes zero sense that using a feature that creates more reads and writes to the nand flash would actually increase a nand flash read and write benchmark performance. Logic dictates that your test is a fluke/invalid.
Looking at the results it's kinda clear that the performance is the same in both on and off benchmarks. I guess benchmark apps are not really a good tool to check if it's bad or good as the memory extension gets used after a certain part of RAM has been used.
THE CPDT BENCHMARK ONLY TESTS 1GB SO IT WONT WORK!!
and remember, Android uses swap file like windows, disk info app says I have 4GB Swap file - so it is already using storage as ram for low priority data.
However, as an example my POCO F3 has UFS 3.1 Flash Storage and LPDDR5 Ram
max theoretical speeds:
LPDDR5 = 6400 Mbps
UFS 3.1 = 2900 Mbps
also:
LPDDR5X=8500Mbps
UFS 4.0 = 5800 Mbps
So, they are all very fast, but the latency will be better with your DDR ram
i think it would be an excellent feature for users with 4GB ram or less
on the funny side, yes you can "download" more ram