Wednesday, January 7, 2009

Memory and Addressing

Memory and Addressing

There are several different types of memory in a micro. One is Program memory. This is where the program is located. Another is Data memory. This is where data, that might be used by the program, is located. The neat (or strange) thing is that they both reside in the same memory space and can be altered by the program. That’s right, a program can actually alter itself if that was necessary. Two terms are used when talking about memory. Reading (load) is getting a value from memory and Writing (store) is putting a value into memory.
There are three buses (not the kind you ride in) associated with the memory subsystem. One is the address bus, the second is the data bus, and the third is the control bus. It's important for you to know exactly how all this works, because these busses transport data and addresses everywhere. All three are connected to the memory subsystem. Its also good to know the function of each to better understand what's happening. In the 8085 CPU, the address bus is 16 bits wide. It acts to select one of the unique 216 (64K) memory locations. The control bus determines whether this will be a read or a write. In the case of an instruction fetch, the control bus is set up for a read operation. Data is read or written through the data bus, which is 8 bits wide. This is why all registers and memory are 8 bits wide, it's the width of the data bus on the 8085 CPU. A bus is just a group of connections that all share a common function. Instead of speaking of each bit or connection in the address separately, for example, all 16 are taken together and referred to simply as the address bus. The same is true for the control and data buses.
A byte is the most used number in a micro because each memory location or register is one byte wide. Memory has to be thought of as a sort of file cabinet with each location in it being a folder in the cabinet. In a file cabinet, you go through the tabs on the folders until you find the right one. To get to each memory location, a different method is used. Instead, a unique address is assigned to each location. In most micros this address is a word or 16 bits, or 4 digit hex. This allows for a maximum of 65536 (216 or 64K) unique addresses or memory locations that can be accessed. These addresses are usually referred to by a 4 digit hex number. Memory usually starts at address 0000h and could go up to FFFFh (216 or 64K or 65536 in total). To access these locations, a 16 bit address is presented to memory and the byte at that location is either read or written.
The Program Counter is what holds this address when the micro is executing instructions. The reason instructions are read sequentially, is because the program counter automatically increments after fetching the current instruction. It does this even before the current instruction is acted upon. The sequence is that the program counter's contents are placed on the memory address bus and the instruction is fetched from memory through the data bus, and immediately the program counter is incremented by 1. Then the micro looks at the instruction and starts processing it. If the instruction is not some kind of jump or call, the instruction is completed and the program counter is presented to the memory address bus again and the next instruction is fetched, the program counter is incremented and the process starts over. This is referred to, in computer jargon, as fetch, decode, and execute.
In the case of reading or writing data, the process is a little different. Data can be read from or written to memory in similar fashion to the fetch. But data does not need the decode & execute steps.

14 comments: