1.4 EXECUTING A PROGRAM

Despite changes in hardware, the basic model of program execution is still that of 'Von Neumann' architecture. John Von Neumann described, in papers published in 1945 and 1946, the basis for a general computing machine, and he is widely regarded as one of its founding fathers. The model relies upon a fetch–execute cycle, where each instruction is copied from its location in RAM into the processing unit, where the operation is carried out before the next instruction is retrieved from memory. A simple program is illustrated in Figure 1.2; it adds together two numeric values stored in memory locations and places the result at a third location.

In earlier days, a program's developer had to write these instructions in binary form. Later, so-called 'assembly languages' allowed the programmer to use simple codes such as LDA, standing for 'load the accumulator'; ADD, 'add two values'; and STO, 'copy a value from the processor's arithmetic unit to a memory location'. A program called an 'assembler' would replace these 'acronyms' by machine-level instructions before execution. So, the program shown in Figure 1.2 can be written as follows:

LDA 6001

ADD 6006

STO 6030

Figure 1.2. A simple program