I believe that the M1 architecture is such that you don't actually need as much RAM as an older model computer. It is a system on a chip so moving data between RAM and storage is so fast that it can swap between the two without much performance degradation.
> It is a system on a chip so moving data between RAM and storage is so fast that it can swap between the two without much performance degradation.
How much faster? I’m not sure what your specific point of comparison is.
Sure, the SoC provides many advantages, many due to Apple’s UMA (Apple’s Unified Memory Architecture) but the system is still limited by storage speed and latency.
But if we’re talking about the relative performance difference between solid state memory (i.e. SSDs, if that is the correct technical term here) and UMA RAM, I’d expect* a significant difference, perhaps (1) an order of magnitude in terms of latency and (2) 2X to 5X in terms of bandwidth. Not to mention that excessive SSD writes would be unwise.
Now I will admit, if there are large parts of the operating system that don’t need to linger / lurk / creep around in RAM, they could be swapped out. And they might even be kept on SSD and not have to be rewritten at all except for patches and upgrades.
* This is a guess based on ‘usual’ data-locality rules of thumb, hopefully allowing for Apple architecture somewhat. I’m happy to be educated / corrected.
