Task Parallelism refers to the distribution of data across multiple cores

User threads are managed with OS support

Kernel threads are supported and managed directly by the OS.

in One-to-One Model Only one user thread can access the kernel at a time.

in Many-to-One Model Threads can run in parallel on multi-core systems.

in One-to-One Model Multiple threads can access the kernel at a given time

OS creates a kernel thread per user thread (can burden performance)

OS creates a set of kernel threads (less than or equal to user threads)

When a thread makes a blocking system call, another thread can run.

Most operating systems use

Fork-join. works well for tasks that are asynchronous (cooperating)

The server creates a thread for each request

in Implicit Threading Programmers identify threads that can run in parallel.

can adjust the number of threads in the pool.

Cancellation of a target thread can be asynchronous or Synchronous.

Implicit fork-join is a synchronous version of the thread pool.

A goal of memory management is not to provide a convenient abstraction for programming

A goal of memory management is to minimize performance with maximum overhead

can be solved by re-mapping between VA
(Virtual Address) and PA (Physical Address)