Yeah, but what I"m inquiring about is a little different.
According to that article the Apple Fusion system fills up the SSD completely 1st. Then, when you're not looking or whatever it moves some off the SSD (to the HDD) until there are 4GB free. Thus there will always be a 4GB free space on the SSD.
Great, so let's say there is exactly 4GB free and a small system file changes. When you're not looking it will then be moved over to the slow HDD. And so on forever until there is a mixture of both frequently and seldomly used large AND small files scattered across both drives. It also means that a 32GB file would write fast for the 1st 4GB and then slow down to HDD speeds for the remaining 28GB.
Well, that sounds pretty dumb (as in not very intelligent) to me and I was wondering what it was like using such a system. Of course if you give it a 250 or 500GB SSD then you will probably never be hitting the HDD at all - no need for a Fusion setup in that case.
Fusion theoretically, should be used to speed up the HDD like a huge cache - only a non-volatile one. Classically and logically you would want something just big enough to fit 100% of all files on the System partition which are smaller than _____ (probably 256K ?) and accessed more than about once a week - plus about 4 to 8GB of managed free space. (I guess that would be about a 32 or 64GB SSD. But to do that an algorithm for sorting and registering candidate files is needed and it SOUNDS LIKE Apple's system doesn't have that.
But I dunno... That's true? Not? Anyone know?
