Thread models and virtual processors

User threads are mapped to kernel threads by the threads library. The way this mapping is done is called the thread model.

There are three possible thread models, corresponding to three different ways to map user threads to kernel threads.

• M:1 model
• 1:1 model
• M:N model

The mapping of user threads to kernel threads is done using virtual processors. A virtual processor (VP) is a library entity that is usually implicit. For a user thread, the VP behaves like a CPU. In the library, the VP is a kernel thread or a structure bound to a kernel thread.

In the M:1 model, all user threads are mapped to one kernel thread; all user threads run on one VP. The mapping is handled by a library scheduler. All user-threads programming facilities are completely handled by the library. This model can be used on any system, especially on traditional single-threaded systems.

In the 1:1 model, each user thread is mapped to one kernel thread; each user thread runs on one VP. Most of the user threads programming facilities are directly handled by the kernel threads. This model is the default model.

In the M:N model, all user threads are mapped to a pool of kernel threads; all user threads run on a pool of virtual processors. A user thread may be bound to a specific VP, as in the 1:1 model. All unbound user threads share the remaining VPs. This is the most efficient and most complex thread model; the user threads programming facilities are shared between the threads library and the kernel threads. This model can be set by setting the AIXTHREAD_SCOPE environment variable to P.