Vous êtes sur la page 1sur 4

EMBEDDED PROCESSOR ARCHITECTURE

Devanshi Sinha Sarthak Seth


Dept. of Electronics and Dept. of Electronics and
Communication, Institute of Communication, Institute of
Technology Technology
Nirma University Nirma University
Ahmedabad, India Ahmedabad, India
16bec033@nirmauni.ac.in 16bec034@nirmauni.ac.in

Further it can be divided on the basis of:


Abstract- Embedded processors are a special
1. Number of instructions executed per clock
type of processors which are designed for a
2. Connection between memory and
special purpose. Embedded systems consist of
processor.
these processors which help them to process the
particular task defined to the system. An
embedded system is very efficient as it focuses to
deliver the assigned task to it in a most efficient II .VON NEUMANN ARCHITECTURE
way, with minimum power consumption. Von Neumann architecture is after the name of the
Today’s era of IOT has a great demand of these
great mathematician John Von Neumann who
embedded processors to fulfil the desires of the thought of a processor whose memory can store
human community, so every processor needs to both data and instruction. Von Neumann
be designed in a special way fulfilling the architecture is one of the most common
expected results, hence we require different architectures that are used in today’s era.
architectures for different hardware designs.
The architectural design decides the properties It is an architecture which is based on a
and functions of that processor.
 Single processor
This paper is about the different  Same memory for both instruction and
architectures which are present and used in data.
these embedded processors.  Single instruction execution at a time.
Keywords: processors, embedded systems, IOT.
The von Neumann architecture consist of three
I .INTRODUCTION
different entities: 1. Processing unit 2.storage unit
Embedded processors can be divided into two (IAS) and 3. Input output unit, which are connected
types: 1. Microprocessors and 2. Microcontrollers. over buses.
Embedded processors are basically used where we
Processing unit- it can be divided into ALU,
require a specific task unlike the utility devices like
control unit, registers, clock.
laptops, and computers.Different processors have
different architectures. Architecture helps us to 1. Arithmetic and logic unit is a part of the
know about the interaction process between a processing unit performs all the arithmetic as
hardware and softwareof the system. It is similar to well as logical operations.
a body skeleton which gives us the outlining and 2. Control unit fetches each instruction that is to
functionality of our whole body. Just by learning be executed and also decodes it for the
the architecture of a given processor we can know hardware.
our processor and its working easily. It can be 3. Registers are the memory circuits which hold
categorized in two basic categories: 1. Von the address of the instruction as well as the
Neumann 2. Harvard architecture. present instruction that is being executed .It
also has the information about the system as the defective programs fail to release
accumulators, status register and interrupt memory.
registers.
4. Clock defines the sped that is faster the clock, Von Neumann bottleneck is a big problem which
more number of instructions per second and can be overcome by some significant changes in
faster the processor. the processor:

 Caching-some part of instructions that are


Immediate access store- it is a memory which used regularly are stored in separate
consist of both program and data and also known as memory locations so that they can be
RAM. Before von Neumann the memory was accessed easily
divided into program and data memory which was  Pre-fetching – it is a technique in which
really complex, but von Neumann proposed that the next instructions are fed into cache
there is no such difference in them and designed a while the present instructions are being
single memory. processed.
 Multithreading – it is a technique of
Input and output unit – any system requires an managing different requests at the same
input and an output port to send and receive data. A time.
port is basically a gateway which allows us to  Advanced RAM – RAM’s like DDR and
control the i/o mechanism. As a user we don’t want SDR can be used which activates both on
to redesign input and output ports hence we use rising and falling edge hence doubling the
different i/o controllers which acts as an interface . speed.

Another architecture was developed to overcome


these problems which were possessed by Von
Neumann, known as Harvard architecture.

IV .HARVARD ARCHITECTURE

Harvard architecture is a design which has two


separate memories for instruction and data.
Harvard architecture got its name from Harvard
Mark 1. In Harvard architecture consist of both
read-only and read-write memory, as the
instructions are only allowed to read whereas the
data can be modified too. As the data and
instruction memory are different their address size
also differs, whereas the instruction address is
wider than data address. Harvard architecture
Fig 1. Von Neumann Model provide us speed but it comes at the cost expensive
processor, more power requirement and
complexity.
III .VON NEUMANN BOTTLENECK

Von Neumann bottleneck is the limitations caused


by this particular architectural design. The term is
on the name of the John Neumann who proposed to
use a same memory for data and instructions. At
the beginning everyone saw the brighter side of it
being simpler and easier to use also it required less
physical space but the era we live in is behind more
and more speed and it is where this architecture
lacked behind. We increased the speed of our
processing unit but the speed of data bus lacked
behind resulting in idle sits of the processor. Many
a times the shared memory leads to crashing of the Fig 2. Harvard Model
V .VON NEUMANN VS HARVARD

Von Neumann is a theoretical design based on the


stored program computer concept which uses same
memory for data and instructions whereas the
Harvard architecture is based on Harvard Mark 1
relay based computer model. It also uses separate
memory for data and instruction. Von Neumann is
slow as compared to Harvard architecture as it
takes one cycle to complete an instruction whereas
von Neumann takes two clock cycles. On the other Fig 3. CISC Architecture
hand Harvard based processors are really costly as
compared to Von Neumann, also it is quite easy
and simple to understand. Harvard architecture instructions were able to support complexity of
comes with a great speed because in this operations. Writing lesser codes that is lesser
architecture data transfer and instructions are number of instructions made programming easier
fetched at the same time hence it is used in for the programmer. Memory cost was one more
microcontrollers and signal processing. factor for the advancement of complex instruction
set. As the cost of memory was high the designers
wanted to decrease the requirement of memory so
VI .MULTI PROCESSOR ARCHITECTURE that cost could be reduced.
Uni-processors cannot provide the performance and
efficiency which is required for the future
applications. Let us think of our cell phones which VIII .RISC
are multi functional and provides us multi In CISC processor the complex instructions take
windows, working on a single processor. The cell large number of processor cycles for execution.
phone will crash immediately with the old set of The slowest operation performed decides the
architectural design, hence to overcome this we use overall speed of the processor in pipelined
multi processor architecture. This has become technology processors. Thus complex instructions
possible with the growing VLSI technology, which reduce the execution speed of even simple
allows us to integrate multi processors on a single instructions. As the compiler technology developed
chip. the programmers started to use higher level
There are a lot of issues which can be easily solved languages. The compiler's work was to convert
with the multiprocessor design, whereas embedding high level codes into machine language codes.
them on a single chip to form a System-On-Chip Combination of simple instructions was used for
also comes with great challenges such as physical achieving complex operations in compilers. It was
and logic design, also we need an architecture seen that writing combination of simple
which is efficient enough for making an instructions rather than one complex instruction
interconnection between processor and peripherals. was much more efficient. Hence compiler
generated programs mostly used simple
instructions. Compilers didn't use most of the
addressing modes offered by CISC processor. The
VII .CISC main aim of product designers became reduction in
CISC is a processor architecture based on read only size and complexity of instruction sets and making
memory and are designed such that it provides simpler and smaller instruction sets that could be
required capabilities in most effective ways. During helpful in 2 ways. First, the speed of pipelining
the development years of computers, coding of could be increased due to simpler instructions and
applications was done through assembly code. Idea hence the performance could be improved. Second,
of high level languages didn't exist at that time. the cost could be reduced as simple instruction set
Writing assembly codes was a tiresome process. requires less computer hardwares. Hence the design
For making programming quick the computers goal was to attain faster execution due to basic
started supporting huge number of instructions. The simple instructions. Compilers used these simple
instructions for construction of complex operations.
REFERENCES

One of the major aspects of this era was increase in


the speed of the processors but still the speed of the
[1]. Ft-sipil.unila.ac.id. (2018). [online] Available at:
memory remained comparatively low. Hence the http://ft-
access of memory became bottle-neck. So the sipil.unila.ac.id/dbooks/The%20Scientist%20and%2
design of registers became one with huge number 0Engineer's%20Guide%20to%20Digital%20Signal%
of registers and cached memory. 20Process.pdf [Accessed 15 Oct. 2018].
[2]. Theteacher.info. (2018). Notes. [online] Available at:
http://theteacher.info/index.php/f453-advanced-
theory/3-3-3-architectures/notes [Accessed 15 Oct.
2018].
[3]. Editor, T., Editor, T. and Editor, T.
(2018). Difference between Harvard architecture
and von-Neumann architecture - Polytechnic Hub.
[online] Polytechnic Hub. Available at:
https://www.polytechnichub.com/difference-harvard-
architecture-von-neumann-architecture/ [Accessed 15
Oct. 2018].
[4]. Anon, (2018). [online] Available at:
https://www.researchgate.net/publication/4053512_
Multiprocessor_architectures_for_embedded_system
-on-chip_applications [Accessed 15 Oct. 2018].
[5]. WhatIs.com. (2018). What is von Neumann
bottleneck? - Definition from WhatIs.com. [online]
Available at:
https://whatis.techtarget.com/definition/von-
Neumann-bottleneck [Accessed 15 Oct. 2018].
[6]. architectures?, W. (2018). What are different types of
computer architectures?. [online] Electrical
Engineering Stack Exchange. Available at:
https://electronics.stackexchange.com/questions/418
Fig 4. RISC Architecture 5/what-are-different-types-of-computer-architectures
[Accessed 15 Oct. 2018].
[7]. En.wikipedia.org. (2018). Harvard architecture.
[online] Available at:
CONCLUSION https://en.wikipedia.org/wiki/Harvard_architecture
[Accessed 15 Oct. 2018].
This paper explains about the embedded processor [8]. En.wikipedia.org. (2018). Von Neumann
architecture. [online] Available at:
architecture. It gives brief introduction of the things
https://en.wikipedia.org/wiki/Von_Neumann_archite
included in embedded processors. It explains about cture [Accessed 15 Oct. 2018].
Von Neumann architecture, Harvard architecture, [9]. Microcontrollers Lab. (2018). WHAT ARE
CISC and RISC architecture. It also includes the EMBEDDED PROCESSORS? Controllers and types.
[online] Available at:
differences between Von Neumann and Harvard
http://microcontrollerslab.com/embedded-processors-
architecture. types/ [Accessed 15 Oct. 2018].
[10]. GHADERI (2018). Risc and cisc. [online]
Slideshare.net. Available at:
https://www.slideshare.net/ghazalghaderiniri/risc-
ACKNOWLEDGEMENT and-cisc-85900408 [Accessed 15 Oct. 2018].
[11]. http://www.romux.com/tutorials/embedded-
system/processor-architecture
This paper has been guided by Prof. Ami Vora. We
are grateful to our teachers for their guidance,
valuable time and relevant sites and materials. We
are obliged to Prof.AmiVora for enhancing our
skills and improving our knowledge about the
topic.

Vous aimerez peut-être aussi