Multicore Programming

Multithreaded Server Architecture

Pros of Multithreading

• Responsiveness - allows for processing to continue even if apart of a process is blocked

• Resource Sharing - threads can easily share data between each other because they are apart of the same process

• Economy - thread switching is cheaper than context switching

• Scalability - can make full use of multicore

Concurrency vs Parallelism

Parallelism - implies a system can perform more than one task simultaneously

Concurrency - supports more than one task making progress (Single processor / core, scheduler providing concurrency)

Data and Task Parallelism

Data parallelism (same op) – distributes subsets of the same data across multiple cores, same operation on each

Task parallelism (diffrent op) – distributing threads across cores, each thread performing unique operation

Amdahl's Law

Performance gains from having additional cores

S serial portion (measure of how much adding extra cores will effect the programs speed)

N processing cores

$speedup=<1/(S+(1-S)/N)$