International Journal of Engineering Research & Technology (IJERT)

ISSN: 2278-0181
Vol. 1 Issue 8, October - 2012

32-Bit Risc Processor For Computer Architecture

Bighneswar Panda. B. Vijay Bhaskar R. Surya Prakash
M.Tech student HOD, Dept. ECE Asst. Prof., Dept. ECE

Avantis St. Theressa Institute Engg & Tech, Chipurupally, Vizianagaram(Dist), AP

Abstract: the time between development and
Now a days computers are used implementation. This way, one can develop a
everywhere in our day to day lives ranging system which can be optimised for manufacturing
from electronics instruments to industrial on a single chip, with the capacity to add or
process automation. Due to complexity of new remove modules according to the requirements in
applications computer engineers use embedded the future. Modern processor design sometimes
systems to develop high performance reduce the implementation effort by acquiring
technological systems which can achieve high some of these elements as Intellectual Property
speed processing while used as hardware (IP) or through implementation techniques to
resources efficiently. To develop embedded build the other components, like VHDL or Verilog
systems, it is necessary to know the basic language and a proprietary synthesizer depending
operation of a computer system, which is on each hardware vendor.
generally composed of memory, a peripheral There exist different approaches to obtain
controller and a microprocessor. This work high comprehension when explaining computer
presents the design and implementation of a 32 architecture. One approach uses FPGA devices
bit RISC processor intended for computer and VHDL language to construct a simple
architecture. This RISC processor should computer, but there is disadvantage that by
RT

support a specific instruction set and its own programming there is a lack of understanding in
assembly code with proper instruction format. dataflow. Another approach uses MSI digital
The instruction set should have simplicity and components such as TTL to construct a computer
IJE

robustness features. It is considered to be an by interconnecting the components but sometimes

effective solution for computer compression. using wires to build those connections is time-
Final circuit can be used as a soft core for consuming.
FPGA embedded designs to control and The design of a 32-bit data width Reduced
automation applications. Set Instruction Computer (RISC) processor was
developed with simplicity and implementation
Keywords: Harvard architecture, PC, Subroutine, efficiency. It has a complete instruction set,
Digital Logic, Microprocessor, ALU, Computer program and data memories, general purpose
Architecture, Programmable Logic, Opcode. registers and a simple arithmetic logical unit
(ALU) for basic operations. It operates following
I. INTRODUCTION a multi-cycle execution nature and is implemented
A digital system was developed by using a on a XILINX Spartan-3e FPGA.
set of interconnected digital integrated circuits like
counters, buffers, logic gates and memory. It II. CPU ARCHITECTURE
requires lots of analysis, testing and to adapt the The processor is based on the Harvard
design metrics like speed, response time, power architecture that any instruction occupies
consumption etc. which resulted very difficult for separated positions of program memory and data
development. Also every design change implied a memory. Thus obtaining greater speed and a
whole analysis and most times expensive to minor program length, also, the access time to the
upgrade. instructions can be superposed with one of the
Now-a-days, advanced technologies like data, obtaining a greater speed in each operation.
programmable logic as Complex Programmable The processor includes a RISC instruction
Logic Devices (CPLD) or Field Programmable set and uses a Single Instruction Single Data
Gate Arrays (FPGA) with more sophisticated (SISD) execution order. Its main characteristics
simulation and design verification environments are:
enable engineers to reach new levels of
complexity and robustness, while greatly reducing
www.ijert.org 1
 International Journal of Engineering Research & Technology (IJERT)
ISSN: 2278-0181
Vol. 1 Issue 8, October - 2012
Eight 32 bit general purpose registers. mnemonic, description, syntax and micro-
256 allocation of 32-bit wide ROM operation done for every instruction.
Program memory The instruction format was developed
256 allocations of 32-bit wide RAM data considering many factors like the CPU
memory architecture, the number of instructions and
ALU with basic arithmetic and logical operands involved. Each instruction has its own
operations. OPCODE (Operation Code), which is analogy to
the assembly mnemonic in machine language.
Program TABLE 1: INSTRUCTION SET
CPU Memory
Mnem
Description Syntax Micro-operation
Registers Data
onic
Memory ADD Addition ADD Rd, Rs RdRd + Rs
SUB Subtraction SUB Rd, Rs RdRd Rs
`Fig 1: The Harvard computer architecture Immediate
ADDI ADDI Rd, k RdRd + k
composed of: Program and Data memories and addition
the CPU interconnected by buses. Immediate
SUBI SUBI Rd, k RdRd k
Subtraction
III. INSTRUCTION SET AND Logic AND AND Rd, Rs RdRd . Rs
To design a CPU, a specific RISC
OR Logic OR OR Rd, Rs RdRd OR Rs
instruction set and its own assembly code with its
proper instruction format is necessary. The NOT Logic NOT NOT Rd RdNOT(Rd)
instruction set should have two objectives: Shift register Rd(n+1) Rd(n),
SHL SHL Rd
RT
simplicity and robustness, i.e. using the simplest left Rd(0) 0
instructions that make the processor capable of Shift register Rd(n) Rd(n+1),
executing complex operations. The instructions SHR SHR Rd
right Rd(7) 0
IJE

are classified into three groups: Immediate

Operations(Arithmetic and Logical) JMP PC PC + k
Jump
Program Control(Jumps) Jump to
JMR JMR Rd PC Rd
Data Manipulation (Load and Storage) register
The arithmetic operations are addition and Branch if if (c=1) then
BRC BRC k
subtraction, because it is possible to perform any carry PC PC + k
other operation with just these two. These
Branch if if (c=1) then
operations can be performed between two BRZ BRZ k
zero PC PC + k
registers and between one register and an
immediate data. Branch if if (c=1) then
BRH BRH k
The logical operations are AND, OR, half-carry PC PC + k
NOT, Shift Right and Shift Left. The first two Load
LDI LDI Rd, k Rd k
operations are executed between two registers, immediate
while the other operations are applied to only one Direct load
register called source register. LDD from LDD Rd,[A] Rd [A]
memory
The branch instructions include
Indirect load
An immediate jump, a jump to a value LDX from LDX Rd,[Rs] Rd [Rs]
contained in a register and memory
Conditional branches Direct
The execution of conditional branches depends on STD storage to STD [A],Rd [A] Rs
the status of the proper flag in the status register. memory
The load and storage instructions required Indirect
must address different data sources on an STX storage to STX [Rd],Rd [Rd] Rs
immediate, direct to data memory and indirect to memory
data memory, addressing modes. Following table Load from
LDP LDP Rd Rd PC
describes the complete instruction set, showing the PC

www.ijert.org 2
 International Journal of Engineering Research & Technology (IJERT)
ISSN: 2278-0181
Vol. 1 Issue 8, October - 2012
Rd: Destination Register; Rs: Source Register; were used, one for the higher word and one for the
k: Constant; PC : Program Counter; A : address lower word (32-bit instruction). The instruction
decoder deal with the raw data stored in the
Instructions are reclassified according to instruction registers and separates it in parts:
type of operands as follows: Operation code (OPCODE), Rd, Rs and A/K, so
Type J (Jumps): only use the OPCODE that these values can be sent to the corresponding
and a constant value (k); component, like the ALU, General Purpose
Type I (Immediate data): use the register block etc. This is achieved simply by
OPCODE, a density (Rd) or source (Rs) using a set of buffers inside a block to sort the
register and a constant value(k); signals to separate buses.
Type R (Register operation): require a General Purpose Resisters (GPR)
single destination register or both: GPRs store and save operands and results
destination and source register. during program execution. ALU and memories
must be able to write/read those registers, so a set
Type J: of eight 8-bit registers were used, along with
OPCODE Not Used K multiplexers and a decoders to control which
register is read or written. Two registers can be
Type I: written at a time.
OPCODE Rd k/A

Type R:
OPCODE Rd Rs Not Used

Fig 2. Instruction format per instruction type.

RT

To specify a register (source or

destination) 3 bits are used since the CPU has 8
general purpose registers. To specify a constant
IJE

value, 8 bits are used. Instruction format is

standardized to 32 bits. Fig. 2 shows the
instruction format for each type.

IV. HARDWARE IMPLEMENTATION

Program Counter (PC)

The program counter produces the address
to read instruction from the program memory. It
can load a random address if the program requires
loops or branches.
ROM Program Memory
The program memory stores the
instructions to be executed. It is to be non-volatile
and fast. It uses internal ROM as program
memory, because it was the fastest option and
eliminated the need for external storage. This
memory was built with 256x1 bit ROM blocks,
with a total of 32 blocks to store 256 32-bit
instructions.
Instruction Register and Decoder
Instruction registers stores the instructions Fig 3. General Purpose Registers Block
read from the program memory and keep it as an
output for the decoder, which separates the
operation code and operands to execute the
command. Two 32-bit registers (D-type flip-flops)

www.ijert.org 3
 International Journal of Engineering Research & Technology (IJERT)
ISSN: 2278-0181
Vol. 1 Issue 8, October - 2012
Arithmetic Logic Unit (ALU)
The ALU has 8 operations; each one of
them was created and converted into a symbol, 0
then a multiplexer was placed to a obtain a 3-bit
selector. Also it has 3 Flags: carry (C), half carry
(H) and zero (Z), which indicate whether there is a
carry, a half carry or a zero after any ALU
operation. 1
RAM Data Memory
The RAM memory is a data storage block,
there the stack is handled and other data are kept
as variables. It is built with 32 memory blocks of
32 bits each one. The address input is divided in 2 2
parts. The first part has 3 bits select the memory
block to read or write using a decoder. The second
has 5 bits that select the memory location between
0 and 31.
3

Fig. 5. Control Unit States Diagram

V. RESULTS
RT

(a) Development Tools

For designing the CPU, Modelsim tool is
IJE

used. Simulation is performed on Modelsim

simulator, designed to work with files generated
by Modelsim.
To simplify the conversion of CPU test
programs, a specific translator is designed in
Verilog HDL language. It is reads .V files which
contains the program in assembly language, and
translates it to machine code, giving the proper
format to load the instructions to the program
memory.
Fig 4. ALU block
(b) Test Application
Control Unit This program is used to evaluate the
The control unit operates as a state performance of the CPU by using all types of
machine. In general, it follows the state diagram instructions to execute the subroutines, access RAM
shown in Fig 5. States 0, 1 and 3 remain equal , and manipulate the stack.
whereas the second state can change depending on
the instruction read by the decoder. In this way:
state 0 represents the fetch/write back stage and
finally during state 3 the program counter is
incremented.

www.ijert.org 4
 International Journal of Engineering Research & Technology (IJERT)
ISSN: 2278-0181
Vol. 1 Issue 8, October - 2012
Table II : TRUTH TABLE FOR CONTROL
UNIT STATE MACHINE

INSTRUCTIONS
GPR_SOURCE1
GPR_SOURCE0
ALU_SOURCE

RAM SOURCE
RAM_WRITE
GPR_WRITE
PC SOURCE
PC_CLOCK

IDR_LOAD
PC_LOAD

IR_LOAD
STATE

0 0 0 0 0 0 0 0 0 0 0 0 Reset
1 0 0 0 1 1 0 0 0 0 0 0 Fetch
JMP,
BRC,
2 1 0 0 0 0 0 0 0 0 0 0
BRZ,
BRH
Fig 6. Simulation of an applicationI (Subroutine jump)
3 0 1 0 0 0 0 0 0 0 0 0 Inc. PC
2 0 0 0 0 0 1 0 0 0 0 0 LDI
2 0 0 0 0 0 1 0 1 0 0 0 LDD
2 0 0 0 0 0 0 0 0 0 1 0 STD
ADDI,
2 0 0 0 0 0 1 1 0 1 0 0
SUBI
ADD,
RT
SUB,
AND,
2 0 0 0 0 0 1 1 0 0 0 0
OR, NOT,
IJE

SHL,
SHR
2 0 0 0 0 0 1 0 1 0 0 1 LDX
2 0 0 0 0 0 0 0 0 0 1 1 STX
2 0 0 0 0 0 1 1 1 0 0 0 LDP
2 1 0 1 0 0 0 0 0 0 0 0 JMR

Fig. 6 shows the jump to the subroutine which is Fig 7. Simulation of a test application (Storage in RAM)
executed when the PC changes from one value to
another value. IV. CONCLUSION
From this task we can observe that
flexibility of a design, which supports a great
modification and improvement of the design. The
simplicity of the design-based on functional blocks
make understanding of each part of a modern
computer works.
The circuit which is obtained can be used as
soft core processor for FPGA designs that would like
to add a CPU to handle other peripheral devices.

www.ijert.org 5
 International Journal of Engineering Research & Technology (IJERT)
ISSN: 2278-0181
Vol. 1 Issue 8, October - 2012
References:
[1] W. Stallings, Organizacin y Arhitecture
Computadoras, 7Editin, Pearson, 2006.
[2] M. Mano, Arquitectura de computadoras,
3 Editin, Pearson, 1994.

[3] A.K.Ray, Advanced Microprocessor And

Peripherals, 2nd Edition, Tata McGraw-
Hill,2008
[4] D.V.Hall,Microprocessor And Interfacing,
2nd Edition, Tata McGraw-Hill,2006
[5] D. Mandalidis, P. Kenterlis, J. Ellinas, A
computer architecture educational system
based on a 32-bit RISC processor,
International Review on computers and
Software, pp. 114-119, 2008.
[6] Tocci, Widmer, Moss, Sistemas Digitales,
Principios y aplicaciones, 10 Editin,Pearson,
2007.
[7] Antonio H. Zavala, Jorge Avante R.,RISC-
Based Architecture for computer Hardware
Introduction
[8] M. Jaumain, M. Osee, A. Richard, A. Vander
Biest, P. Mathys, Educational simulation of
RT

the RiSC processor, ICEE International

Conference on Engineering Education, 2007.
IJE

[9] P. Verplaetse, J. Campenhout, ESCAPE:

Environment for the Simulation of Computer
Architecture for the Purpose of Education,
IEEE TCCA Newsletter, February, pp. 57-59,
1999.
[10] G. Wainer, S. Daicz, L. De Simoni, D.
Wassermann, Using the Alfa-1 simulated
processor for educational purposes, Journal
on Educational Resources in Computing, vol.1,
Issue 4, pp.111151, 2001
[11] J. Djordjevic, A. Milenkovic, N. Grbanovic,
An Integrated Environment for Teaching
Computer Architecture, IEEE Micro vol.20,
Issue 3, pp. 6674, 2000.
[12] M. Becvar, A. Pluhacek, J. Danecek, DOP: a
CPU core for teaching basics of computer
architecture, Workshop on Computer
architecture education, Article No. 4, 2003.
[13] H. Elaarag, A complete design of a RISC
processor for pedagogical purposes, Journal
of Computing Sciences in Colleges, vol. 25,
Issue 2, pp. 205-213, 2009.

www.ijert.org 6