Lecture # 04

Vector and Array Processor

PARALLEL AND DISTRIBUTED COMPUTING
CS-482
Types of Array Processor
Array Processor performs computations on large array of data. These are two
types of Array Processors: Attached Array Processor, and SIMD Array
Processor. These are explained as following below.
1. Attached Array Processor :
2. SIMD array processor :
1. Attached Array
Processor :
To improve the performance of the host computer in numerical computational
tasks auxiliary processor is attached to it.
Attached array processor has
two interfaces:
1.Input output interface to a common processor.

2.Interface with a local memory.

Here local memory interconnects main memory. Host computer is general

purpose computer. Attached processor is back-end machine driven by the
host computer.

The array processor is connected through an I/O controller to the computer &
the computer treats it as an external interface.
Working:
Let us assume that we are executing vast number of instructions, in that case
it is not possible to execute all instructions with the help of host computer.
Sometimes it may take days of weeks to execute these vast number of
introductions. So, in order to enhance the speed and performance of the host
computer as shown in above diagram.

I/o interface is used to connect and resolve the difference between host and
attached process. Attached array processor is normally used to enhance the
performance of the host computer . Array processor mean bunch/group of
process used together to perform any computation.
2. SIMD array processor :
This is computer with multiple process unit operating in parallel Both types of
array processors, manipulate vectors but their internal organization is
different.
SIMD array processor
SIMD is a computer with multiple processing units operating in parallel.

The processing units are synchronized to perform the same operation under
the control of a common control unit. Thus, providing a single instruction
stream, multiple data stream (SIMD) organization. As shown in figure, SIMD
contains a set of identical processing elements (PES) each having a local
memory M.
Working:
Here array processor is in built into the computer unlike in attached array
processor where array processor is attached externally to the host computer.
Initially mark control unit decodes the instruction and generate the control
signals and passes it into all the processor elements(PE’s) or ALU upon
receiving the control signals from master control unit, all the processing
elements come to throw that operations need to perform. The data perform
the operations will be accessed from main memory into the local memory of
respective PE’s. SIMD array processor is normally used to compute vector
data. All PE’s execute same instruction simultaneously on different data.

for ex: Ci= Ai + Bi

Here the vector instruction Ci= Ai+ Bi need to be executed is addition

operation, the master control unit generate control signal and passes it onto
all processing elements and data.
 Working:
Parallelly on same instruction on with different data, that’s why it is called as
Single instruction multiple data streams(SIMD) array processor.

Each PE includes –

• ALU

• Floating point arithmetic unit

• Working registers
Working:
Master control unit controls the operation in the PEs. The function of master
control unit is to decode the instruction and determine how the instruction to
be executed. If the instruction is scalar or program control instruction then it
is directly executed within the master control unit.

Main memory is used for storage of the program while each PE uses
operands stored in its local memory.
Operation of SIMD Array
Processor
The SIMD form of parallel processing is called Array processing. Figure
shows the array processor. A two-dimensional grid of processing elements
transmits an instruction stream from a central control processor. As each
instruction is transmitted, all elements execute it simultaneously. Each
processing element is connected to its four nearest neighbors for the
purposes of data exchange. Connections around the end can be provided on
both rows and columns, but they are not shown in figure.
Vector processor
classification
Vector processors have rightfully come into prominence when it comes to
designing computing architecture by virtue of how they handle large datasets
efficiently. A large portion of this efficiency is due to the retrieval from
architectural configurations used in the implementation. Vector processors
are classified into two primary architectures: memory to memory and register
to register. These classification are important to optimize performance on the
scientific computing and other data intensive applications.
What is Vector Processor
Classification?
Vector processor classification is the labeling of processors based on the
amount of vector operations it could be handling at a time. Here are the main
classifications of vector processors:
Memory to Memory
Architecture
In memory-to-memory architecture, source operands, intermediate results,
and final results are retrieved directly from the main memory. For memory-to-
memory vector instructions, it is necessary to specify the base address,
offset, increment, and vector length to facilitate data transfers between the
main memory and pipelines. Notable processors employing memory-to-
memory formats include TI-ASC, CDC STAR-100, and Cyber-205.

Main points about Memory-to-Memory Architecture

• No limitation on size.

• Speed is comparatively slower.

 Register to Register
Architecture
In register-to-register architecture, operands and results are retrieved
indirectly from the main memory through the use of a large number of vector
or scalar registers. Processors like Cray-1 and Fujitsu VP-200 utilize vector
instructions in register-to-register formats.

Main points about Register-to-Register Architecture

• Limited size.

• Higher speed compared to memory-to-memory architecture.

• Increased hardware cost.

Hybrid Architecture
Hybrid Architecture unites memory-to-memory and register-to-register
architectures to gain from all. The way is that of flexible operand retrieval
methods; which improves performance and efficiency in many computational
tasks. This solution provides a balanced solution which can be adapted to
requirements of the application, with a better utilization of available
resources.
A block diagram of a modern multiple
pipeline vector computer is shown below:
Advantages of Vector
Processor
• Parallelism and SIMD Execution: Vector processors are intended to perform Single Guidance,
Different Information (SIMD) activities. This implies that a solitary guidance can work on numerous
information components in equal, considering a huge speedup in errands that include monotonous
procedure on enormous datasets. This parallelism is appropriate for logical recreations, reenactments
of actual frameworks, signal handling, and different applications including weighty mathematical
calculations.
• Proficient Information Development: Vector processors are enhanced for moving information
productively among memory and the computer chip registers. This is pivotal for execution, as memory
access is much of the time a bottleneck in many registering errands. Vector processors commonly
incorporate specific information development directions that can move information in huge lumps,
limiting the effect of memory dormancy.
• Diminished Guidance Above: With SIMD tasks, a solitary guidance can play out similar procedure
on various information components. This lessens the above related with getting, unraveling, and
executing individual directions, which is especially helpful for errands that include tedious
computations.
• Energy Proficiency: By executing procedure on various information components at the same time,
vector processors can accomplish higher computational throughput while consuming generally less
power contrasted with scalar processors playing out similar activities consecutively. This energy
effectiveness is significant for superior execution registering (HPC) applications where power
utilization is a worry.
 Advantages of Vector
Processor
• Logical and Designing Applications: Vector processors succeed in logical and designing
reproductions where complex numerical calculations are performed on huge datasets. Applications like
weather conditions displaying, computational liquid elements, atomic elements reproductions, and
picture handling can benefit altogether from the computational power and parallelism presented by
vector processors.
• Superior Execution for Cluster Activities: Some true applications include controlling exhibits or
grids of information, for example, in information examination, AI, and illustrations handling. Vector
processors can perform activities like expansion, increase, and other component wise procedure on
whole clusters with a solitary guidance, significantly speeding up these undertakings.
• Compiler Streamlining: Vector processors frequently expect code to be written with a certain goal in
mind to make the most of their capacities. Compiler improvements can naturally change undeniable
level code into vectorized guidelines, permitting software engineers to zero in on the algorithmic plan
as opposed to low-even out advancements.
• Memory Transmission capacity Usage: Vector processors can productively use memory transfer
speed by streaming information from memory into vector registers and afterward performing
calculations on those vectors. This can assist with alleviating the exhibition effect of memory
bottlenecks.
• Vector Length Adaptability: Some vector processors permit adaptability in the length of vector
tasks, empowering software engineers to pick the proper vector length in light of the application’s
necessities and the accessible equipment assets.
Disadvantages of Vector
Processors
• Complexity of Programming: Development of such codes is complex
because optimizing code for vector processors generally requires
specialized knowledge.

• Cost: Vector processors are expensive, because they require special

hardware.

• Limited Applicability: Not all applications are good candidates for vector
processing; there are tasks that are inherently sequential and cannot
profitably be executed in parallel.
Limitations of Vector
Processor
•Speed Limitation: Vector processors are limited by the speed at which they
can execute instructions.

•Memory Limitation: Vector processors are limited by the amount of

memory available for storing data.

•Instruction Limitation: Vector processors can only execute certain

instructions and often require instruction modifications for more complex
tasks.

•Cost Limitation: Vector processors are often more expensive than scalar
processors due to their advanced technology.
Conclusion
Vector processors provide an important means of improving computational
efficiency, for example, in applications where lots of math and big data
abound. If developers understand their classifications and architectural
configurations, then they can use their capabilities effectively to realize
better performance in different computational tasks.
What is the Primary Advantage
of Using Vector Processors?
Operations on large datasets at the same time, much faster than with regular processors, are
what vector processors are good at, for use in scientific and engineering datasets.
How does memory-to-memory architecture
differ from register-to-register architecture?

The difference in performance characteristics affects the architectures of memory to memory

and register to register memory architecture in which data retrieval is done directly from main
memory and to registers respectively.
Vector Processor vs Scalar
Processor
A processor is an essential component of a computer system, responsible for
carrying out instructions in order to facilitate various computer operations.
Traditionally, processors have been either vector processors or scalar
processors, both of which have their own unique set of benefits and
drawbacks. Vector processors are designed to complete multiple data
operations in one instruction and can provide higher performance than scalar
processors. Conversely, scalar processors are designed to carry out one
instruction at a time and are more efficient for certain types of operations. In
this article, we will discuss the differences between vector processors and
scalar processors and how they are used in modern computing.
Vector Processor
A vector processor is a type of processor that can process multiple data
elements at once. It is capable of performing operations on a vector of data
elements in parallel. Vector processors are particularly useful for tasks such
as image and video processing, where large amounts of data need to be
processed in parallel. Vector processors are also used in scientific computing,
where they are used to accelerate the processing of complex algorithms.
Characteristics of Vector
Processor
• Vector Processors are designed to process multiple data elements in
parallel, while Scalar Processors process one data element at a time.
• Vector Processors can be more efficient, as they can complete a given task
with fewer instructions than a Scalar Processor.
• Vector Processors are more complex than Scalar Processors, and require
more memory and power to operate.
• Vector Processors are used for more demanding tasks, such as scientific
calculations and 3D game rendering, while Scalar Processors are used for
simpler tasks, such as basic calculations and web browsing.
• Vector Processors are more suitable for data-intensive applications, while
Scalar Processors are better suited for applications that require fewer
calculations.
• Vector Processors can be more expensive than Scalar Processors, as they
require more complex hardware and software.
Applications of Vector
Processor
• Computer-Aided Design (CAD): CAD software allows for the creation of realistic 3D
models, which can be used for product design, engineering, and architecture. Vector
processor power makes it easier to manipulate complex models and make changes
quickly.
• Image Processing: Vector processors are used to manipulating and analyzing images.
This can include tasks such as edge detection, object recognition, and facial recognition.
• Virtual Reality: Vector processors are used to rendering realistic 3D graphics in virtual
reality applications. This allows users to experience a more immersive experience when
interacting with virtual worlds.
• Scientific Computing: Vector processors are used to performing complex calculations in
scientific computing applications. This can include tasks such as calculating weather
patterns or complex simulations.
• Artificial Intelligence: Vector processors are used to helping train and run neural
networks for artificial intelligence applications. This can include tasks such as facial
recognition, object recognition, and natural language processing.
• Data Analysis: Vector processors are used to analyzing large amounts of data quickly.
This can include tasks such as analyzing customer data or financial data.
Scalar Processor
A scalar processor is a type of processor that can process one data element
at a time. Scalar processors are typically used for general-purpose computing
tasks, such as word processing and spreadsheets. Compared to vector
processors, scalar processors are less powerful and slower, but they are
cheaper and more energy-efficient.
Characteristics of Scalar
Processor
• Scalar processors deliver high-performance capabilities, capable of executing
multiple instructions simultaneously.

• Scalar processors execute a single instruction in one clock cycle.

• Scalar processors are relatively low-cost compared to other types of processors.

• Scalar processors have low power consumption, making them more efficient and
cost-effective.

• Scalar processors are highly flexible and can be used to solve a variety of problems.

• Scalar processors are relatively simple in design and are easy to program and
maintain.
Advantages of Scalar
Processor
• Low Cost: Scalar processors are typically much cheaper than vector
processors, making them more accessible to many people.
• Low Power Consumption: Scalar processors are much more efficient
than vector processors, reducing the amount of power needed to operate
them.
• Easier to Program: Scalar processors are simpler to program than vector
processors, making them easier to use for novice programmers.
• Flexible: Scalar processors are more flexible than vector processors,
allowing them to be used in a variety of applications.
• High Clock Speed: Scalar processors are able to process instructions at a
much faster rate than vector processors, increasing the speed of
computations.
• Good for Single-Threaded Tasks: Scalar processors are better suited for
single-threaded tasks, as they can process one operation at a time.
Limitations of Scalar
Processor
• Memory Bandwidth: Scalar processors are limited by the amount of
memory bandwidth available.

• Instruction Set: Scalar processors typically have limited instruction sets and
lack features found in modern processors.

• Performance: Scalar processors are often limited in their performance due

to their single-instruction execution.

• Compatibility: Scalar processors are often limited in their compatibility with

vector processors due to their different instruction sets.

• Processing Speed: Scalar processors are typically slower than vector

processors.
Applications of Scalar
Processor
• Video Games: Scalar processors are used to rendering 3D graphics for video
games. This allows for faster and more realistic graphics, which can enhance the
gaming experience.
• Web Browsing: Scalar processors are used to rendering webpages quickly and
accurately. This allows for a smoother and faster browsing experience.
• Office Applications: Scalar processors are used to rendering documents and
spreadsheets quickly. This allows for faster and more efficient document editing.
• Graphics Applications: Scalar processors are used to rendering graphics
quickly. This can include tasks such as image manipulation and video editing.
• Database Applications: Scalar processors are used to quickly search and
query large databases. This can include tasks such as customer data retrieval or
financial data analysis.
• Mobile Applications: Scalar processors are used to rendering graphics quickly
for mobile applications. This allows for a smoother and faster user experience
when interacting with mobile apps.
Vector Processor v/s Scalar
Processor
Factor Vector Processor Scalar Processor
It has a single instruction multiple data (SIMD) architecture. It has a single instruction single data (SISD)
Architecture Instructions are executed multiple times in parallel on
multiple data elements.
architecture. Instructions are executed one at a
time on a single data element.

Execution It can execute instructions faster than a scalar processor due It takes more time to execute instructions since the
Time: to the nature of SIMD architecture. instructions are executed one at a time.

Data It processes multiple data elements in parallel. It processes single data element at a time.
Processing
Registers It has multiple registers to store multiple data elements.
It has a single register to store a single data
element.

Cache It has multiple caches for multiple data elements. It has a single cache for the single data element.

Branching It does not support conditional vector branching. It supports conditional vector branching.
Vector Processor v/s Scalar
Processor
Factor Vector Processor Scalar Processor

Performance It offers higher performance than a scalar processor due to It offers lower performance than vector processors
its SIMD architecture. due to its SISD architecture.

Power It consumes more power than a scalar processor due to its It consumes less power than a vector processor
Consumption more sophisticated architecture. due to its simpler architecture.

Cost It is more expensive than a scalar processor due to its

sophisticated architecture.
It is cheaper than a vector processor due to its
simpler architecture.

Parallelism It has more parallelism due to multiple data elements being It has less parallelism due to a single data element
processed in parallel. being processed at a time.

Memory It requires more memory bandwidth due to multiple It requires less memory bandwidth due to single-
Bandwidth memory accesses. memory access.
Vector Instruction Format in
Vector Processors
Vector instruction format is a type of instruction format used in vector processors, which are
specialized types of microprocessors that are designed to perform vector operations efficiently. In a
vector processor, a single instruction can operate on multiple data elements in parallel, which can
greatly accelerate certain types of computations, such as matrix multiplication, signal processing,
and scientific simulations.

Vector instruction format is a type of instruction format used in vector processors, which are
specialized processors designed to perform operations on vectors and matrices efficiently. In vector
processors, a single instruction operates on multiple data elements in parallel, which is different
from scalar processors where a single instruction operates on a single data element at a time.

The vector instruction format typically includes an opcode field that specifies the operation to be
performed, a vector register specifier field that identifies the vector register that contains the data
to be operated on, and a length field that specifies the number of elements in the vector.

The vector processor reads the vector instruction and retrieves the data from the vector register
identified in the instruction. It then performs the specified operation on all the elements in the
vector in parallel, and stores the result back into the same vector register or a different one,
depending on the instruction.
Vector instruction formats can vary depending on
the specific vector processor architecture, but
they typically include some combination of the
following fields:
1. Opcode field: specifies the operation to be performed on the vector data.

2. Vector register specifier field: identifies the vector register that contains the data to be
operated on.

3. Length field: specifies the number of elements in the vector.

4. Mask field: specifies which elements in the vector to operate on.

5. Addressing mode field: specifies how to access the vector data in memory.

6. Data type field: specifies the data type of the elements in the vector.

Vector instruction formats allow vector processors to perform operations on large amounts of
data in parallel, making them well-suited for scientific and technical applications, such as
simulations, modeling, and image processing.
Vector instruction set
Different Instruction formats are used by different vector processors.
Vector instructions are generally specified by some fields. The main fields
that are used in vector instruction set are given below:
Operations Code (Opcode)
Base addresses
Operations Code (Opcode)
1.Operations Code (Opcode) – The operation code must be specified to
select the functional unit or to reconfigure a multi-functional unit to perform
the specified operation dictated by this field. Usually, microcode control is
used to set up the required resources. For example: Opcode – 0001
mnemonic – ADD operation – add the content of memory to the content of
accumulator Opcode – 0010 mnemonic – SUB operation – subtract the
content of memory to the content of accumulator Opcode – 1111
mnemonic – HLT operation – stop processing
 Base addresses:
Key addressing components
1. Base addresses – For a memory reference instruction, the base addresses are needed for both source
operands and result vectors. The designated vector registers must be specified in the instruction, if the operands
and results are located in the vector register file, i.e., collection of registers. For example: ADD R1, R2. Here, R1
and R2 are the addresses of the register.
2. Offset (or Displacement) – This field is required to get the effective memory address of operand vector. The
address offset relative to the base address should be specified. Using the base address and the offset (positive
or negative), the effective address is calculated.
3. Address Increment – The address increment between the scalar elements of vector operand must be
specified. Some computers, i.e., the increment is always 1. Some other computers, like TI-ASC, can have a
variable increment, which offers higher flexibility in application. For example:
R1 <- 400
Auto incr-R1 is incremented the value of R1 by 1.
4. Vector length – The vector length (positive integer) is needed to determine the termination of a vector
instruction.