How to program multicore

Question

4.00/5 (3 votes)

See more:

Is it appropriate to have a program that uses a core to do an intensive calculation and another core to output the results simultaneously on a dualcore machine? can i use MPI library to do this task?

I tried multithreading on a dualcore machine with the same idea that one thread does the calculation while the other thread displaying results, but the two cores do not process simultaneously.

Thanks,

Posted 10-May-11 15:35pm

ktn0805

Add a Solution

2 solutions

Solution 1

"but the two cores do not process simultaneously" ??

Well, depending on the "particular" of each thread there might be a delay of some kind between the thread that does the calculation and the thread that display the results.

IMO, you should get a good experience base using threads and all the synchronisation mechanisms before using something more complex (like mpi) unless you absolutly need features that it offers.

Anyway, good luck

M.

Posted 10-May-11 15:54pm

Maximilien

Comments

Sergey Alexandrovich Kryukov 10-May-11 22:00pm

It has nothing to do with delay. In fact, all cores do work "simultaneously", but the scheduler does not have to distribute the thread on one process onto difference cores.

Please see my answer.
--SA

Add a Solution

Add your solution here

Treat my content as plain text, not as HTML

Preview 0

…

Existing Members

Sign in to your account

...or Join us

Download, Vote, Comment, Publish.

Your Email
Password
Forgot your password?

Your Email
This email is in use. Do you need your password?
Optional Password

I have read and agree to the Terms of Service and Privacy Policy
Please subscribe me to the CodeProject newsletters

When answering a question please:

Read the question carefully.
Understand that English isn't everyone's first language so be lenient of bad spelling and grammar.
If a question is poorly phrased then either ask for clarification, ignore it, or edit the question and fix the problem. Insults are not welcome.
Don't tell someone to read the manual. Chances are they have and don't get it. Provide an answer or move on to the next question.

Let's work to help developers, not make them feel stupid.

This content, along with any associated source code and files, is licensed under The Code Project Open License (CPOL)

Sergey Alexandrovich Kryukov · Accepted Answer · 2011-05-10T15:58:00

Short answer: you do not control it explicitly. The system runs several threads on several cores. Which core is loaded with which thread at each thread time spice is decided by the system threading scheduler.

However, there are ways to affect this behavior programmatically.
Remember, in almost all cases this is not recommended! The cases when it can help are very rare.

See the following Windows API: GetThreadAffinityMask/SetThreadAffinityMask, GetProcessAffinityMask/SetProcessAffinityMask and SetThreadIdealProcessor.

General information on thread affinity: http://en.wikipedia.org/wiki/Processor_affinity[^],
see also this discussion: The perils of thread affinity:
http://www.bluebytesoftware.com/blog/PermaLink,guid,8c2fed10-75b2-416b-aabc-c18ce8fe2ed4.aspx[^].

On MSDN, start here:
http://msdn.microsoft.com/en-us/library/ms685096(v=VS.85).aspx[^],
http://msdn.microsoft.com/en-us/library/ms684251(v=vs.85).aspx[^].

Again, don't interfere in thread/process affinity unless you know exactly what are you doing!

(I know just one case when modification of affinity can be helpful: working with standard hardware (typically industrial) which is "bad" in the sense that it requires extensive polling as interrupt-driven operation is node accessible via API; in this case it is beneficial to dedicate on code to some threads serving these "bad" parts of equipment, which makes other threads' timing to be more predictable.)

[EDIT]

Parallel computing is a big topic which goes well above the usage of CPU/cores.
For approaches to parallel computing, see:
http://technet.microsoft.com/en-us/library/ff919485(v=ws.10).aspx[^],
http://www.microsoft.com/downloads/en/details.aspx?FamilyID=f9d74066-0dfd-4a38-9bad-bff5e15c9629[^],
http://academic.research.microsoft.com/Publication/1693102/on-task-mapping-optimization-for-parallel-decoding-of-low-density-parity-check-codes-on[^].

It will give you the idea on what's involved. Do your research, find some more articles.

Start with: http://en.wikipedia.org/wiki/Parallel_computing[^].

—SA