
Chapter 2
Chapter 2
Architecture
2.1 Architectural Overview
2.1.1 Overview
Architecture
The EM60000 series have a high performance 8-bit RISC CPU with the
Havard-type architecture which allows the program and data to have separate
memories in order to improve the bandwidth of data access. Applying a
two-level overlapped fetch-and-execute pipeline, all instructions will only
require a single instruction cycle (except for those that modify the program
counter). Note that each instruction cycle takes two system clocks, resulting in
a high instruction throughput and fast interrupt response.
Arithmetic Logic Unit (ALU)
The EM60000 series contain an 8-bit ALU and accumulator (ACC). The ALU
is a general-purpose arithmetic unit that performs 2’s-complement arithmetic
and logical operations. It operates under all of the data memory in direct or
indirect address mode that includes memory-mapped control registers. The
ALU also supports BCD decimal arithmetic operations.
Data Memory
The data memory of EM60000 series has a total of 144 bytes with direct and
indirect address modes. Three locations of the data memory are set aside to be
shared by the multiplier. All operations can be applied directly to the data
memory. Part of the control registers are mapped to the data memory and other
part are mapped to the IOC register (I/O mapped control register). The data
memory is divided into one common and four banked blocks selected by 2
control bits.
Hardware Multiplier
The 8x8 hardware multiplier of EM60000 series takes only a single instruction
cycle. It performs signed multiplication and generates a 16-bit signed result that
is stored in two registers mapped in the data memory. The multiplication
instruction can be applied on data memory or on an immediate constant,
providing a wide range of user applications, such as Digital Signal Processing.
EM60000 Series User’s Manual
Architecture
•
5