CS 3060 Class Notes 02/03/16 - Ch 04 Threads
THREAD LIBRARY with this we are talking about:
APIs
END OF CHAPTER 3
END OF CHAPTER 3
Socket two parts
IP address (identifies the computer); Port number (identify that particular program)
OLE
Object Link Embed
Benefits of a thread? SLIDE 4-4
Responsiveness; Resource Sharing; Economy; Scalability;
START OF CHAPTER 4
START OF CHAPTER 4
THREADS
THREADS
POSIX is a
an API
Which mode does the multithreading happen in?
both; user and kernel mode;
What are the two types of parallelism:
data; task;
Data parallelism:
distributes subsets of the same data across multiple cores, same operation on each.
Task parallelism:
distributing threads across cores, each thread performing unique operation.
One to one is where:
each user level thread maps to kernel thread; creating a user level thread creates a kernel thread;
OLE example:
excel being opened inside of word is an example of this.
XDL is:
external data representation
Process creation is:
heavy weight
Many to many thread what is happening?
many kernel threads to many user threads as long as the number of kernels and users are different numbers otherwise it would be one to one
There is such a thing as fake parallelism which is referred to as: SLIDE 4-14
many-to-one; point of this is fairness; everybody gets a chance to run;
What does the thread library provide?
provides the programmer with API for creating and managing threads.
Light weight thread creation does what?
saves us many resources
Multiple tasks with the application can be implemented by:
separate threads;
Threading is better because you can:
share the code, share the data (global data), share the local data (the stack),
As the number of threads grows what happens?
so does architectural support for threading.
Multicore can only only work if you upgrade your _____.
software - otherwise it can't keep up and you won't see a difference anyway.
AMDAHL'S LAW
speed up formula:
Pthreads may be provided at what level?
user level and kernel level;
Which modes?
user mode; kernel mode;
If you increase your numbers of CPU
you increase the speed
Multicore systems putting pressure on programmers, the challenges include:
■ Dividing activities; ■ Balance; ■ Data splitting; ■ Data dependency; ■ Testing and debugging - this is head ache for a multi thread;
