Exploring the World of
Underlying and Parallel
Computing
This presentation delves into the fascinating worlds of underlying and
parallel computing, exploring their types, challenges, and applications.
by Abhi Yadav
Underlying Computing: A Deeper Look
Definition Key Concepts
Underlying computing involves distributing computational Resource sharing, task distribution, and load balancing are
tasks across multiple interconnected computers, often crucial elements of underlying computing.
referred to as a "cluster" or "grid."
Types of Underlying
Computing
Cloud Computing Grid Computing
Utilizes remote servers to Leverages geographically
deliver on-demand computing dispersed computers to solve
resources, such as storage, complex scientific problems.
databases, and software.
Cluster Computing
Employs a group of tightly connected computers working together to
enhance performance.
Challenges and Limitations of Underlying
Computing
1 Network Latency 2 Data Management 3 Security
Communication delays between Ensuring data consistency and Protecting data and resources
nodes can impact performance. integrity across distributed from unauthorized access.
systems.
Applications of Underlying
Computing
Web Services Scientific Research
Powering large-scale web Accelerating simulations and
applications and platforms. analyses in fields like genomics and
physics.
Financial Modeling
Performing complex financial
calculations and risk assessments.
Parallel Computing: The
Power of Simultaneous
Execution
Parallel computing involves dividing a task into smaller
1
parts, each of which can be executed concurrently on
multiple processors.
This approach aims to improve overall performance by
2
leveraging the processing power of multiple CPUs.
Parallel computing is widely used in fields that require
3
massive computational power, such as scientific modeling
and data analysis.
Types of Parallel Computing
SIMD (Single Instruction, Multiple Data)
1 Executes the same instruction on multiple data elements
simultaneously.
MIMD (Multiple Instruction, Multiple Data)
2 Allows multiple processors to execute different instructions
on different data sets concurrently.
Data Parallelism
3 Distributes data across multiple processors and operates on
it concurrently.
Challenges and Limitations of Parallel Computing
1 Synchronization
2 Communication Overhead
3 Load Balancing
4 Programming Complexity
Applications of Parallel
Computing
1 2
AI/ML Image Processing
Training large language models Real-time image recognition and
and deep learning algorithms. computer vision tasks.
3
High-Performance
Computing
Modeling climate change,
simulating molecular dynamics.
Underlying vs. Parallel Computing: Key Differences
Underlying
1
Focuses on task distribution across multiple computers.
Parallel
2 Emphasizes simultaneous execution on multiple processors within a single
computer.
Key Distinction
Underlying involves distributing tasks across a network,
3
while parallel computing focuses on concurrent execution
within a single system.
