My searches on SO have failed me, so if this is a duplicate, please redirect me.
With that out of the way, my question: I learned, from experience and browsing SO that a Java boolean is stored as a 32-bit int if you declare it as a standalone value, but as an 8-bit byte if you declare it within an array. My question, here, is as follows: Which is more memory efficient? Does the meta data of the array make it bigger in memory than the alternative?
boolean oneVariable = false, oneArray[] = {false};
The “meta data” of the array includes:
Add on the 1 necessary byte for the boolean data and you have 13 bytes (32 bit) or 21 bytes (64 bit) at a minimum.
However, objects are allocated memory in 8-byte multiples, so even though you only need 12 or 20 bytes of overhead + 1 byte for the boolean, you’ll end up using 16 or 24 bytes of memory, respectively, for your array object.
In addition to the 16/24 bytes the object itself will take up, you’ll need 4 bytes (32 bit) or 8 bytes (64 bit) for the memory address of the object, totaling 20 or 32 bytes of memory, respectively, to store your boolean in an array.
The size of a standalone variable is JVM dependent. Java does not specify the size of storage, and in fact Oracle says
Older JVMs use a 32-bit stack cell, used to hold local variables, method arguments, and expression values so a single boolean used as a variable would consume 4 bytes; making the array at least 5 times as expensive as for a single boolean. This answer may be different if, for example, the boolean is a class variable in which case it would just be a single byte added to the existing overhead. In newer JVMs a single boolean would only use 1 byte, but depending on its context and the 8-byte padding necessary to align memory addresses, could still consume up to 8 bytes of heap space. It would still be smaller than the boolean array.