"The influence of noise is just your assumption". Yes, I arrived at this guess after all other guesses of - files being different, defragging, and other causes you listed were found to be not true. The other system noises causing changes to sq gives me more confidence in this approach. If you do have other suggestions, do let me know, I’ll check that too! I don't think neglecting this case scenario would be a good idea. I donot have the necessary equipment to measure the access noise, else I would have tried that too.
Regarding the explanations, I have mentioned it multiple times. I don’t know much about modern HDDs (the head alignment etc involves a lot of modern control theory, and almost each batch has its own specific firmware), but with flash storage it is stored as a set of charges within a specific threshold value inside a floating gate NAND transistor cell. The pattern can change the noise profile when accessing. Modern devices are far more complex than the basic structures I have seen in academic resources and I am searching for leads to check the actual pathway/design (need to check everything from firmware etc, and I am not sure how successful it will be considering that most of these are kept as secrets).
Regarding the explanations, I have mentioned it multiple times. I don’t know much about modern HDDs (the head alignment etc involves a lot of modern control theory, and almost each batch has its own specific firmware), but with flash storage it is stored as a set of charges within a specific threshold value inside a floating gate NAND transistor cell. The pattern can change the noise profile when accessing. Modern devices are far more complex than the basic structures I have seen in academic resources and I am searching for leads to check the actual pathway/design (need to check everything from firmware etc, and I am not sure how successful it will be considering that most of these are kept as secrets).