You are on page 1of 6

The 14th International Conference on

Computer Science & Education (ICCSE 2019)


August 19-21, 2019. Toronto, Canada WedP1.4

Multi-Cycle CPU Design With FPGA for Teaching


of Computer Organization Principle
Xing Liu Linghao Fu Wenbi Rao
Computer Science and Technology School Computer Science and Technology School Computer Science and Technology
Wuhan University of Technology Wuhan University of Technology Wuhan University of Technology
Wuhan, China Wuhan, China Wuhan, China
liu.xing@whut.edu.cn fulinghao@whut.edu.cn wbrao@whut.edu.cn

Xiongmin Lin * Mingxi Liao Bing Shi


Electrical and Computer Engineering Computer Science and Technology School Computer Science and Technology School
University of Victoria Wuhan University of Technology Wuhan University of Technology
Victoria, Canada Wuhan, China Wuhan, China
xiongminlin@uvic.ca cs lmx@whut.edu.cn bingshi@whut.edu.cn

Abstract—CPU design with programmable FPGA has become experiment platform are solidified and cannot be expanded or
a significant solution for the experiment teaching of computer reprogrammed. Therefore, the students have few opportunities
organization principle. In this paper, the design concept, im- to redesign the experimental programs independently by their
plementation technique and verification method of a multi-
cycle MIPS CPU are presented. To decrease the CPU design own knowledge. In most situations, they can only follow
complexity, the hierarchical division approach which divides the the instructions of the pre-designed experiment cases, and
CPU into a serial of easy-to-implement modules is proposed. To observe the executing results. On the other hand, the hands-on
help the students master the intricate FPGA development process wire connection is tedious and difficult to debug, especially
in an easy-to-understand way, the easy-to-hard progressive imple- when the physical damage occurs. As a result, not only is
mentation technique is investigated. To verify the functionality of
each CPU module, the variable-control approach which verifies the experiment teaching efficiency decreased, but also the
the CPU modules progressively by comparing with the standard students’ experimental interests are reduced.
ones is explored. The experiment cases have been carried out To address the above challenges, it is significant to imple-
for six years, and the course survey results showed that the ment an experiment program which has the following features:
new experiment design with the new implementation approaches (1) the hardware of the experiment platform can be extended
had improved the teaching quality of the computer organization
experiment significantly. or reprogrammable flexibly, so that the students can have
Index Terms—Computer organization, experiment, CPU, enough free space to design the experiment from their own
FPGA. viewpoints. (2) a user-friendly development tool needs to be
selected, so that the students can keep away from the tedious
I. I NTRODUCTION hands-on connection operations and implement the complex
Computer organization principle is accepted as one of the circuits more efficiently.
key hardware courses in the computer science curricula [1], One solution to satisfy the above requirements is to require
[2]. To improve the teaching quality of computer organization, the students to develop a MIPS CPU with the reprogrammable
a decent experimental design is significant as it can deepen FPGA. With the reconfigurability of FPGA, the students
the students’ understanding to the theoretical knowledge and can design their own experimental cases independently and
strengthen their ability in applying what they have learned. flexibly by their own knowledge. Moreover, they can design
Traditional computer organization experiments are con- the circuit system by drawing the schematic diagram with
ducted on the integrated experiment platform which has so- the FPGA development tool, rather than build the circuit
lidified all the chip elements such as controller, arithmetic system by the hands-on wire connection. Currently, the above
logic unit (ALU), memory and I/O devices. The students only experiment design concept has been researched and realized
need to configure the circuit manually by connecting the chips by many institutions. Taking the computer organization exper-
with wires, and then verify the functionality of the circuits iment course as an example, Lee et al. [3] presented the work
by uploading a program and controlling the I/O devices of designing a five-stage pipelined 32-bit MIPS CPU which
such as switches and buttons. This experimental scheme is supported 21 MIPS instructions on the Altera DE2 FPGA
conductive to improve the students’ hands-on skills. Yet, it board, for the purpose of teaching the computer architecture
has several limitations: On the one hand, the chips on the more effectively. Wang et al. [4] proposed a method to design
the data path and controller of a multi-cycle MIPS CPU on the
*Corresponding author: Mingxi Liao. Email: cs lmx@whut.edu.cn Nexys-4 FPGA board, for the purpose of reducing the CPU

978-1-7281-1846-8/19/$31.00 ©2019 IEEE 472


Authorized licensed use limited to: Gheorghe Asachi Technical University of Ia¿i. Downloaded on January 16,2024 at 09:23:11 UTC from IEEE Xplore. Restrictions apply.
WedP1.4

design and implementation difficulty. Qin et al. [5] investigated III. CPU D ESIGN BY H IERARCHICAL D IVISION S CHEME
a method on how to design both the static and dynamic
pipelined CPU on the Nexys-4 FPGA board by improving a The objective of the experiment is to enable the students to
single-cycle CPU, while Wan et al. [6] explored a simulation design and implement a multi-cycle MIPS CPU independently
methodology which can help the students understand the MIPS based on the SWORD FPGA platform. Since the multi-cycle
system more easily. CPU architecture has a certain complexity, it is difficult for
In this paper, we present the experiment teaching of com- some students to complete this work. To solve this problem,
puter organization on how to design and implement a multi- the hierarchical division approach which divides the CPU from
cycle MIPS CPU with FPGA platform. This paper is organized top to bottom into a series of easy-to-implement modules is
as follows: In Section II, the lab environment is introduced. In applied. Then, these modules can be implemented and verified
Section III, the design concept of a multi-cycle MIPS CPU is one by one. And when all the modules pass the verification
presented. In Section IV, the CPU technical implementation test, they can be assembled to build a completed CPU. In
scheme is explained. In Section V, the CPU verification Figure 2, the project overview for this experiment is shown.
mechanism is investigated. In Section VI, the application Project 1 and 2 require students to integrate a SOC system
development for the designed CPU is presented. In Section by using the CPU Intellectual Property (IP) cores. This SOC
VII, the course survey and evaluation work are introduced. system can provide a debugging and evaluation environment
Finally in Section VIII, the conclusion to this paper is given. for the CPU development later. Although the objective of the
experiment is to develop a CPU. Yet, if there is only one
CPU, without any other auxiliary peripheral devices, the CPU
II. L AB E NVIRONMENT
cannot operate effectively. To provide a decent development
To serve the experimental teaching, the SWORD (Simple environment for CPU, the CPU needs to be embedded into
While Organic aRc Design) board is developed and used as the a SOC system. This SOC system, depicted as it is shown in
experimental platform for the computer organization teaching, Figure 3, consists of the core CPU as well as the peripheral
depicted in Figure 1. SWORD is equipped with Xilinx Kintex- components such as the ROM, RAM, Multiple IO (MIO) bus,
7 XC7K160T-1FFG676 FPGA, 512M bytes DDR3 memory, counter, seven-segment LEDs, buttons and switches. To build
32M bytes NOR FLASH and 6M bytes SRAM memory. It this SOC system, the auxiliary elements such as the counter,
has abundant peripheral interfaces such as the SPI, VGA, clock division, MIO bus, mechanical anti-jitter are developed
HDMI, SFP+, MicroSD, RS232, SATA, Pmod and USB, as in project 1 firstly. Then, in project 2, these elements are
well as a set of GPIO interfaces such as the seven-segment integrated with the CPU, ROM and RAM so that a complete
LEDs, buttons and switches. SWORD can connect to the host SOC system can be built.
machine by JTAG interface, and can be developed by the From project 3 to project 8, the experiment focuses on
software such as the ISE design suite 14.7 or the Vivado design the multi-cycle MIPS CPU development. Since the CPU
suite. architecture has a certain complexity, the hierarchical division
In the following sections, we will present our computer development approach is applied. First, in project 3, the CPU
organization experiment work on how to design and implement is divided into two modules: CPU controller and CPU data
a multi-cycle MIPS CPU with the SWORD FPGA platform path. The controller is responsible for parsing the instruction
and the ISE design suite. opcode, while the data path is responsible for parsing the
instruction address. Then, in project 4 and 5, the experiment
focuses on designing and implementing the CPU data path
and CPU controller respectively. In project 4, the hierarchical
division approach is applied further. With this approach, the
RS232 USB-OTG/HID
VGA CPU data path is divided into a set of easy-to-implement ele-
MicroSD Ethernet
ments, including the Arithmetic Logical Unit (ALU), register
set and transfer circuits. These elements are implemented and
SATA Power
interface
verified one by one, and then integrated to build the data
NOR SPI
HDMI path. Similar to project 4, project 5 also uses the hierarchical
PmodTM Arduino division approach to develop the CPU controller. It divides
Extension
FPGA SRAM
SFP+
interface
the controller into a series of elements such as the opcode
SDRAM
parser and the function code parser. Then, it implements these
HDMI elements and integrate them to build the CPU controller.
SATA
interface Seven-Segment LEDs On completion of project 5, the students are capable of
Buttons
USB JTAG
building a single-cycle MIPS CPU. Next, in project 6, the
students are required to extend the architecture of the CPU
Switches
controller and CPU data path so as to make the CPU support
more MIPS instructions. After this extension, the students
Fig. 1. The SWORD FPGA board for MIPS CPU design and implementation. also need to make assembly programs in which the extended

473
Authorized licensed use limited to: Gheorghe Asachi Technical University of Ia¿i. Downloaded on January 16,2024 at 09:23:11 UTC from IEEE Xplore. Restrictions apply.
WedP1.4

Project 2 CPU Project 4 Extension Project 6 Project 8


CPU Design of single-cycle Extension of CPU Multi-cycle MIPS
Build SOC by IP cores Datapath
division CPU Datapath instructions CPU Design
CPU (IP core) division
Extension of CPU Multi-cycle CPU
Project 3 Transfer Circuit
RAM (IP core) Datapath Datapath Design
Project 1


Build CPU by IP cores
Early-stage Preparation ROM (IP core) Registers
Extension of CPU Multi-cycle CPU
CPU Datapath Controller Controller Design
MIO Bus MIO Bus Built by IP cores ALU
 
Counter Counter CPU Controller Project 7
Built by IP cores Project 5 CPU interruption
Clock Division Clock Division
Design of single-cycle Illegal instruction
Anti Jitter Anti Jitter interruption
CPU Controller Extension
CPU Arithmetic overflow
Seven-segment LED Seven-segment LED Opcode parsing interruption
Controller
Buttons/ Switches Buttons/ Switches division FunCode parsing External interruption

Fig. 2. Project overview for multi-cycle MIPS CPU design.

       

5'< FON
6HYHQ'LJLW6HYHQ6HJPHQW/(''LVSOD\
5671 5671
8
&5 FON
0$L  UHDGQ UHDGQ UVW UVW
8
&5
VHJBFON VHJBFON
%71B\ 
FON .H\BRXW  'LQ 
$L 
'LY  6WDUW D D
.H\B\  .H\BUHDG\ 'BUHDG\
6:B2.  6: VHJBVRXW VHJBVRXW I E I E
.H\B[ 
SXOVHBRXW  3XOVH  %71B2.  %L  %L  J J
%71B2.  %71 
6:  6:B2.  6:B2.  6:B2.  6:B2. 
'LY  IODVK 6(*B3(1 6(*B3(1 H F H F
&WUO  EOLQN  EOLQN 
UHDGQ
6:B2. 
UVW
UVW +H[V 
VHJBFOUQ VHJBFOUQ
G GS G GS
$
6$QWLBMLWWHU 6(QWHUBB SRLQW  $1 $1 $

%71B[  /(6 


66HJB'HY
9&&

FON FONGLY  'LY  /(BRXW 
8 FONBGLY
6: 6: 6: 6:
6:  UVW
SRLQWBRXW  9
&ONB&38 ,2BFON
6:B2.  6: &ONB&38
'LVSBQXP 
,19
1
FONBPK] FON
UVW 'LVSBQXP 
*1'
(1 SRLQWBRXW 

7HVW  /(BRXW 

'LY  'LY  SRLQWBLQ 

^ E` /(6 


FON
&38,2 
% UVW 8 3HULSKHUDOBLQ 
*3,2HBZH 113& 
'DWD 
GDWD 
8 %

*3,2)

0XOWLB&+
%71  *3,2IBZH LQVW  GDWD 
6:  &RXQWHUBRXW  GDWD 

FON

UHVHW
8 PHPBZ PHPBZ 0,2B%86
$GGUBRXW 
'DWDBRXW 
GDWD 
GDWD 
$GGUBRXW 
$GGUBRXW  DGGUBEXV  'DWDBLQ  GDWD 

LQVW  LQVWBLQ  'DWDBLQ  3&  GDWD 
'DWDBLQ  &SXBGDWDEXV 
/('BRXW 
'DWDBRXW  OHGBRXW 
'DWDBRXW  &SXBGDWDEXV 

,17 6&38 FON FRXQWHUBVHW 


&38B0,2 UDPBGDWDBLQ 
FRXQWHUBRXW 
&RXQWHUBRXW 
UVW 8 /('BRXW 
*3,2I 
3&  3&BRXW  0,2BUHDG\ GDWDBUDPBZH FRXQWHUBRXW (1 OHGBFON OHGBFON
FRXQWHUBRXW
UDPBDGGU  OHGBVRXW OHGBVRXW
FRXQWHUBRXW 'LY  6WDUW
63,2 OHGBFOUQ /('B3(1
/('B3(1
& UDPBGDWDBRXW  FRXQWHUBZH &
3B'DWD  OHGBFOUQ
/('

3& 
D 
8 VSR 
FON FRXQWHUB287

9*$
520B' UVW
'LY 
FON
&RXQWHUB[
FRXQWHUB287
'LY 
FON 8
'LY 
FON
DGGUD 

ZHD 
8GRXWD  FRXQWHUBZH
FRXQWHUB287

'
GLQD 
FRXQWHUBYDO 

FRXQWHUBFK 
FRXQWHUBRXW  .(<%2$5' '

FOND 5$0B%
,19

       

Fig. 3. Schematic of the SOC system (Partial).

instructions are used, and then run the programs on the buffer. In Figure 4, the data path of the extended multi-cycle
extended CPU so as to verify whether the extension is correct. CPU is shown. In addition to the datapath, the CPU controller
Project 7 extends the functionality of CPU further by also need to be redesigned, and this can be achieved by
enabling the CPU to process the interrupts such as illegal designing an instruction state machine and converting this state
instructions, arithmetic overflow and some other external in- table into the VHDL programming. In Figure 5, the instruction
terrupts. This extension involves the work of extending the state machine for the extended CPU controller is shown.
CPU controller by adding some elements such as the interrupt
detection circuit, anomaly detection circuit and interrupt re- IV. CPU D EVELOPMENT BY E ASY- TO - HARD
sponse control circuit. It also involves the work of extending I MPLEMENTATION T ECHNIQUE
the CPU data path by adding some registers and instruction Different technical solutions can be applied to implement
paths. the CPU modules, including the IP core reuse, the schematic
Finally, in project 8, the students are required to extend the design and the VHDL programming. IP core reuse is the
controller and datapath of the single-cycle CPU so that a multi- simplest way to implement the modules. Yet, not every module
cycle CPU can be built. To perform this extension, the CPU can have the IP core. If not, the students should develop the
datapath needs to be extended by adding the Program Counter modules independently by VHDL programming.
(PC) transfer path, the register set transfer path as well as the In order to make it easier for the students to master the
transfer path between the instruction register and the memory CPU implementation technique with FPGA. We have designed

474
Authorized licensed use limited to: Gheorghe Asachi Technical University of Ia¿i. Downloaded on January 16,2024 at 09:23:11 UTC from IEEE Xplore. Restrictions apply.
WedP1.4

       

FON
UHVHW FON 3&
UVW

0,2BUHDG\ 4  3&B&XUUHQW 

5(*
&(
$ $
3&:ULWH
$1' ' 

3&:ULWH&RQG
25
%UDQFK $1'
$1'
3&6RXUFH 
3& V 

, 

%UDQFKBSF R 
, 

-XPSBDGGU

08;7 B
-XPSBDGGU 3&B&XUUHQW  ,QVW  11 , 

1RW8VH
, 

,RU'
% $/8B2XW  %

$/8BRSHUDWLRQ 

$/86UF%  FON


V
$/82XW , 

$/86UF$

08;7B
1 UVW R 
0BDGGU 
4  , 

5HJ:ULWH
9

5(*
&(

5HJ'VW  V
' 
,5
, 

FON /B6

08;7B
FON R 

8 UGDWDB$  $ 


, 

UHVHW UVW
,QVW 
UVW $/8BRSHUDWLRQ 

4  UV ,QVW 


& 5BDGGUB$  &
,5:ULWH
5HJV ]HUR ]HUR
5(*

&(
UW ,QVW 
5BDGGUB% 

8
DOX
' 
V 
UHV 
, 
UHV 
08;7B
UG ,QVW 
, 
:WBDGGU 
0HPWR5HJ  1RW8VH
R 

, 
RYHUIORZ RYHUIORZ
1RW8VH V 

1
, 

FON UGDWDB%  , 


R 

0'5 V 
,  %  9&&

08;7 B
$/823
1 UVW ,  :WBGDWD  , 

0'5 R 

4  ,  , 

RIIVHW
08;7B

9
5(*

&( 1RW8VH , 

Imm_32 9
1RW8VH ,  ,PPB  *1'
GDWDBRXW 
GDWD&38  ' 
Imm_16 ([WB
LPPB 
' '

,QVW 

       

Fig. 4. Schematic of the multi-cycle CPU data path.


.2314$3*/.'&3$)
.2314$3*/.%&$/%&
1&(*23&1'&3$)
complexity of the VHDL programming can be simplified.
&-&"% 
 1$
/1  1$
However, from project 5, the schematic design technique is
3"13 !1*3&
 1$
 1$
 0 no longer allowed to be used. Instead, the students should
 0
!1*3&
/41$&
&
implement the CPU controller completely by the VHDL pro-
360



0

 gramming. Therefore, the implementation difficulty is further




0


&-/16"%%1&22 
!
0

0 1".$) 4-0




$/-043"3*/. 1

0

!/
5&$43*/.

$/-0,&3*/.

$/-0,&3*/.
improved if compared with the project 4.
 1$
 1$
 1$
 1$
 1$
 1$
 0
!1*3& Through the above progressive implementation mechanism,
 0 /41$&
!1*3&/.%
 0 
/41$& the students can understand the functionalities of the IP core,
schematic design and VHDL programming more deeply, and

0!

0



master the CPU implementation technique more easily.


!


&-/16 &-/16
"$$&22 "$$&22 360&$/-0,&3*/.


&(23
&-&"%
/1
&-!1*3&
/1
&(!1*3&
&-3/&(
V. CPU V ERIFICATION BY VARIABLE - CONTROL
M ECHANISM
!1*3&#"$+23&0
With the hierarchical division approach, a complex CPU can
&(23  
&(!1*3&
be divided into a set of easy-to-implement modules. These
&-3/&(  
 modules can be developed one by one, and then assembled
to build a complete CPU. However, before being assembled,
the functionality of each module needs to be verified. In this
Fig. 5. State machine design for CPU controller development.
section, we present two kinds of verification approaches. One
is the verification by the ISE simulation, and the other is
different levels of implementation strategies for the CPU the verification by the real-world variable-control testing. The
development, as it is shown in Table I. former approach can be used to verify the functionality of each
In project 3, the CPU is divided into two modules, and the independent module in the CPU, while the latter one can be
IP cores of these two modules are provided to the students used to verify the systematic functionality of the entire CPU.
directly. The students can load these IP cores to build a 1) Verification by simulation: Simulation is a common
CPU rapidly. This project is easy to be implemented, and its mechanism to verify the functionality of an independent
objective is to perform the first division of the CPU and enable module in the CPU. In Figure 6, a simulation to verify the
students to learn how to integrate a CPU with the IP cores. opcode parsing operations in the CPU controller is shown. By
From project 4, the IP cores are no longer allowed to be setting a serial of input values and checking the correctness
used. The students should remove the IP cores provided in of the output, whether the functionality of the register set is
project 3 one by one, and then develop these modules inde- correct or not can be determined.
pendently by VHDL programming. Compared with project 3, 2) Verification by variable-control approach: Variable-
the development difficulty is improved significantly. control approach refers to the strategy which substitutes only
In project 4, the schematic design technique is still allowed one of the many parameters in the system each time, and
to be used. With the support of the schematic diagram, the then check whether this substitution has changed the original

475
Authorized licensed use limited to: Gheorghe Asachi Technical University of Ia¿i. Downloaded on January 16,2024 at 09:23:11 UTC from IEEE Xplore. Restrictions apply.
WedP1.4

 
              

   



         
   
 !     
    "    
   #     
 $     

Fig. 6. Verify the CPU opcode parsing circuit by ISE simulation.

functionality of the system, so as to judge whether this the students. In Figure 8, some application examples such as
substitution is correct or not. snake game, Tetris game and ping-pong game are shown.
In Figure 7, the verification process of the variable-control
mechanism is shown. Provide the system is composed of three
modules: A, B and C. First, the standard IP cores of these VII. C OURSE S URVEY AND E VALUATION
modules are loaded, and a standard-compliant system is built.
Suppose a test program T is input to this system, and its output Before 2014, we conducted the computer organization ex-
is R. Next, the standard IP core of module A is removed, and periment on the traditional solidified experiment board which
the students are required to develop this module independently used the hands-on connection. Yet, from 2015, we started to
by themselves. Then, to check whether the development of do experiments by designing a multi-cycle MIPS CPU on
module A is correct or not, the test program T is input the SWORD FPGA board. In 2017, we started to apply the
again to the system, and its output O1 is compared with the new teaching methods such as variable-control verification and
standard output R. If O1 is matched with R, it indicates analogous teaching to our teaching work.
the module A has been developed properly. Likewise, the To evaluate the teaching effect, we did two kinds of
following modules such as B and C can be developed and survey to the students. One is whether the new experiment,
verified in the same way until the entire system is built by the which design a multi-cycle CPU with FPGA, has helped
students independently. you understand the computer organization principle better
Currently, the above variable-control approach has been if compared with the old experiment which performs the
applied in our CPU development work. First, the standard IP hands-on connection on the solidified board. The survey result
cores of different CPU modules are provided to build a CPU. showed that almost all the students thought the new experiment
Then, these IP cores are removed one by one. Once removed, was more contributive. Precisely, 36% students thought the
the students need to develop this module independently by new experiment was more contributive in a small degree, while
Verilog programming, and then verify whether this substitution 62% thought it was more contributive in a large degree.
is correct by using the variable-control approach. The other survey is whether the variable-control verification
and analogous teaching methods have helped you conduct
VI. A PPLICATION D EVELOPMENT ON CPU the experiment better? We did this survey in the year 2016
After the CPU is built, the students are also encouraged to and 2017, and the survey results are shown in Figure 9
develop some applications programed by assembly language, and Figure 10, respectively. From the results, it is shown
and then run these applications on the pre-built CPU. By doing that the new teaching method applied in 2017 has improved
this, the students can understand how do the CPU operate more the teaching quality positively if compared with the teaching
deeply. Currently, many applications have been developed by situation in 2016.

476
Authorized licensed use limited to: Gheorghe Asachi Technical University of Ia¿i. Downloaded on January 16,2024 at 09:23:11 UTC from IEEE Xplore. Restrictions apply.
WedP1.4

Test Stage 1 Standard output R


Program T Module A Module B Module C
built by standard IP core built by standard IP core built by standard IP core

Stage 2 Module A
Module A Module B Module C R = O1 ?
Output O1 is built correctly
built by students built by standard IP core built by standard IP core

Stage 3
Module B
Module A Module B Module C R = O2?
Output O2 is built correctly
built by students built by students built by standard IP core

Stage 4 Module C
Module A Module B Module C R = O2?
Output O3 is built correctly
built by students built by students built by students

Fig. 7. Verification process by using variable-control approach.

                

Fig. 8. Application development by assembly on the pre-built CPU.

result showed that the new experiment design with the new
teaching approach had improve the teaching quality of the
computer organization experiment significantly.
ACKNOWLEDGMENT
This research is funded by National Natural Science
Foundation of China (Grant No. 61702387), Natural Sci-
ence Foundation of Hubei Province of China (Grant No.
Fig. 9. Survey on how do the students master each project (in 2016). 2017CFB302), Computer Education Research Association of
Chinese Universities (CERACU2019R07), and Cooperation
Education Project of Huawei and Ministry of Education of
China (201802001024).
R EFERENCES
[1] IEEE Computer Society and ACM, Curriculum Guidelines for
Undergraduate Degree Programs in Computer Engineering, 2016.
[Online]. Available online at: http://www.acm.org/binaries/content/
assets/education/ce2016-final-report.pdf. [Accessed January 2019].
[2] ACM/IEEE-Joint Task Force on Computing Curricula. Computer
Science Curricula 2013 Final Report [EB/OL]. (2013-12-20).
Fig. 10. Survey on how do the students master each project (in 2017). http://ai.stanford.edu/users/sahami/CS2013/. [Accessed January 2019].
[3] Lee, J. H., Lee, S. E., Yu, H. C., & Suh, T. (2012). Pipelined cpu design
with FPGA in teaching computer architecture. IEEE Transactions on
Education, 55(3), 341-348.
VIII. C ONCLUSION [4] Wang, L., Yu, Z., Zhang, D., & Qin, G. (2018, August). Research on
Multi-Cycle CPU Design Method of Computer Organization Principle
The new computer organization experiment which target to Experiment. In 2018 13th International Conference on Computer Science
build a multi-cycle MIPS CPU with FPGA is presented in & Education (ICCSE) (pp. 1-6). IEEE.
[5] Qin, G., Hu, Y., Huang, L., & Guo, Y. (2018, August). Design and
this paper. In order to enable the students to develop the CPU Performance Analysis on Static and Dynamic Pipelined CPU in Course
better, the hierarchical division approach and the easy-to-hard Experiment of Computer Architecture. In 2018 13th International Con-
progressive implementation technique are proposed. To verify ference on Computer Science & Education (ICCSE) (pp. 1-6). IEEE.
[6] Wan, H., Gao, X., & Long, X. (2016). Simulation Methodology Used
the functionality of each CPU module, the variable-control in Computer Structure Course. In Theory, Methodology, Tools and
approach which verifies the CPU modules progressively by Applications for Modeling and Simulation of Complex Systems (pp.
comparing with the standard ones is investigated. The survey 237-244). Springer, Singapore.

477
Authorized licensed use limited to: Gheorghe Asachi Technical University of Ia¿i. Downloaded on January 16,2024 at 09:23:11 UTC from IEEE Xplore. Restrictions apply.

You might also like