A branch table is a method to efficiently transfer program control from one part of a program to another, or to a second program that has been loaded dynamically, by sending a branch of instructions or by implementing a switch statement. This method, sometimes called a jump table, relies on a set of circumstances or conditions to perform a jump to implement a procedure according to a switch statement from a C+ compiler. Some of the main advantages of branch tables are their compact code structure and a reduction in the need to test return codes individually when determining program flow.
In the 1980s, branch tables were used widely in assembly language programming. They are still used in assembly programming for embedded systems and operating systems development. Since the 1990s, compiler programming languages have also made use of branch table functioning.
Branch tables consist of a list of unconditional instructions that, given input, branch off to other destinations. Most computer hardware can execute these instructions efficiently. At times, the offset — essentially, the distance to the destination — can be added to a program counter register that can then point to sets of branch instructions or jump among sets of branch instructions. All that is necessary to implement a branch table is validation of the input code, transforming the data into an offset, and multiplying the data to a given instruction length.
Embedded programming uses branch tables as they are more memory efficient than using machine code or arrayed pointers. Embedded control systems need that memory savings, and though it may cost a small amount of performance during access to the branch table, any virtual method function call would cost the same amount of performance for stable functioning. The limited CPU access and memory savings in embedded systems need a branch table for static function sets.
Compiler programming languages generate their own branch tables when needed by use of search keys on optimizing compilers. Some programmers choose to manually assist the compiler with contexted branch table generation by giving the compiler two-step conditional parameters from which to search keys. In the earlier years of compiling programs, the branch table implementation used to be the "GoTo" commands in Fortran compiler languages. Branch tables are still used to implement changes in program flow in compiler languages or to be a starting point for repeated instruction sequences.