Table of Contents
Advertisement
1
1.1 Outline of the 32170
1.1.1 M32R Family CPU Core
(1) Based on RISC architecture
• The 32170 is a 32-bit RISC single-chip microcomputer which is built around the M32R family
CPU core (hereafter referred to as the M32R) and incorporates flash memory, RAM, and
various other peripheral functions-all integrated into a single chip.
• The M32R is based on RISC architecture. Memory access is performed using load and store
instructions, and various arithmetic operations are executed using register-to-register
operation instructions. The M32R internally contains sixteen 32-bit general-purpose registers
and has 83 distinct instructions.
• The M32R supports compound instructions such as Load & Address Update and Store &
Address Update, in addition to ordinary load and store instructions. These compound
instructions help to speed up data transfers.
(2) 5-stage pipelined processing
• The M32R uses 5-stage pipelined instruction processing consisting of Instruction Fetch,
Decode, Execute, Memory Access, and Write Back. Not just load and store instructions or
register-to-register operation instructions, compound instructions such as Load & Address
Update and Store & Address Update also are executed in one cycle.
• Instructions are entered into the execution stage in the order they are fetched, but this does not
always mean that the first instruction entered is executed first. If the execution of a load or
store instruction entered earlier is delayed by one or more wait cycles inserted in memory
access, a register-to-register operation instruction entered later may be executed before said
load or store instruction. By using "out-of-order-completion" like this, the M32R controls
instruction execution without wasting clock cycles.
(3) Compact instruction code
• The M32R instructions come in two types: one consisting of 16 bits in length, and the other
consisting of 32 bits in length. Use of the 16-bit length instruction format especially helps to
suppress the program code size.
• Some 32-bit long instructions can branch directly to a location 32 Mbytes forward or backward
from the instruction address being executed. Compared to architectures where address space
is segmented, this direct jump allows for easy programming.
1-2
OVERVIEW
1.1 Outline of the 32170
Ver.0.10
Advertisement
Table of Contents
