sizeof : Java Glossary


Java does not have a sizeof operator like that in C/C++. All primitive types have a standard size (or at least appear to have a standard size), so you don’t really need one. There are typically no pad or alignment bytes.
Java Object Size
type size in bytes
overhead ~8 bytes/object
boolean 1
byte 1
char 2
short 2
int 4
long 8
float 4
double 8
reference 4/8†
String length * 2 + 4
† addresses are 4 bytes in a 32-bit JVM (Java Virtual Machine) and 8-bytes in a 64-bit JVM.

A JVM is free to store data any way it pleases internally, big or little endian, with any amount of padding or overhead, though primitives must behave as if they had the official sizes. For example, the JVM or native compiler might decide to store a boolean[] in 64-bit long chunks like a BitSet. It does not have to tell you, so long as the program gives the same answers. It might allocate some temporary Objects on the stack. It may optimise some variables or method calls totally out of existence replacing them with constants. It might version methods or loops, i.e. compile two versions of a method, each optimised for a certain situation, then decide up front which one to call. Then, of course, the hardware and OS (Operating System) have multilayer caches, on chip-cache, SRAM (Static Random Access Memory) cache, DRAM (Dynamic RAM) cache, ordinary RAM (Random Access Memory) working set and backing store on disk. Your data may be duplicated at every cache level. All this complexity means you can only very roughly predict RAM consumption.

There is a header on each object, typically 8 bytes for objects and 12 bytes for arrays and 16 bytes in JET.

In Java version 1.5 or later you can use java.lang.instrument. Instrumentation. getObjectSize()

Other methods you may find useful in estimating RAM use:

Minimum and Maximum Possible Primitive Values

Primitive variables include Boolean, char, byte, short, int, long, float and double. Strings, arrays and Objects are not considered primitives.
Java Primitives
Type Signed? Bits Bytes Digits Lowest Highest Mnemonic
Boolean n/a 1 1 1 false true zero/one
char unsigned Unicode 16 2 4:5 '\u0000' [0] aka Character. MIN_VALUE '\uffff' [216-1] aka Character. MAX_VALUE Unicode chars are twice as big as C’s.
byte signed 8 1 2:3 -128 [-27] aka Byte. MIN_VALUE +127 [27-1]aka Byte. MAX_VALUE Bytes are signed, so half the usual 255 range.
short signed 16 2 4:5 -32,768 [-215] aka Short. MIN_VALUE +32,767 [215-1] aka Short. MAX_VALUE 32K
int signed 32 4 9:10 -2,147,483,648 [-231] aka Integer.MIN_VALUE aka -2 gig, roughly -2 billion +2,147,483,647 [231-1] aka Integer. MAX_VALUE. aka 2 gig, roughly 2 billion 2 gig
long signed 64 8 1:19 -9,223,372,036,854,775,807 [-263] aka Long. MIN_VALUE about -9×1018 9,223,372,036,854,775,808 [+263-1] aka Long. MAX_VALUE about 9×1018 9 exabytes, or 9 billion gig
float signed exponent and mantissa 32 4 7 ±1.40129846432481707e-45 aka Float. MIN_VALUE ±3.40282346638528860e+38 aka Float. MAX_VALUE
or roughly ±2127
with about 7 significant digits of accuracy.
A float can exactly represent integers
in the range -224 to +224.
rough, compact float
double signed exponent and mantissa 64 8 16 ±4.94065645841246544e-324 aka Double. MIN_VALUE ±1.79769313486231570e+308 aka Double. MAX_VALUE
or roughly ±21023
with 15 significant digits of accuracy, almost 16 with 15.95 significant digits.
A double can exactly represent integers
in the range -253 to +253.
high precision float

This page is posted
on the web at:

Optional Replicator mirror
on local hard disk J:

Canadian Mind Products
Please the feedback from other visitors, or your own feedback about the site.
Contact Roedy. Please feel free to link to this page without explicit permission.

Your face IP:[]
You are visitor number