- Parallel computer - memory and clock is shared
- Distributed computer - memory and clock is dedicated
- In Distributed, there are two types: Master and Slave
- E.g, master is main node and if there is any update, then master gets the update first and then the slave gets it.
- High computation power is used in parallel computing
- Whereas, distributed does not have high computation power.
- Parallel - solving one large problem and doing things simultaneously.
- Serial Computing, breaks the problem into the stream of instructions, executed one at a time.
- In Parallel computing, problem is broken into parts and solve the problems concurrently.
- Used for weather modeling, automated stock trading, database, web based services, search engines, transaction processing
- Parallel programming - single source, single memory address and the connected ones are tasks.
- Hybrid model
- e.g, two clusters that are connected and have their own CPU but memory is shared.
- OpenMP
- Message Passing
- MPI
- Message Passing Interface - messages between the tasks
- MPI + OpenMP = Hybrid Model
- Individual clusters are made using OpenMP
- To connect the clusters, MPI is used.
- Platform As A Service = PaaS
- Software As A Service = SaaS
- Infrastructure As A Service = IaaS
- Distributing Computer System
Recent Pastes