Thread Parallel Efficient

Question

1.67/5 (3 votes)

See more:

Write a parallel function that will multiply two n by n matrices with a variable number of threads.
This should be an efficient by choosing the number of threads and number of blocks for each thread.

Thanks in advance!

Posted 5-Mar-12 9:04am

Mhd Sami Almouhtasb

Updated 7-Mar-12 2:19am

Manfred Rudolf Bihy

v3

Add a Solution

Comments

Sergey Alexandrovich Kryukov 5-Mar-12 16:02pm

Very good, but not a question.
Do you just want to order that someone do your job for your? Good luck, then.
Or do it by yourself and ask a question when and if you stuck.
--SA

Mhd Sami Almouhtasb 5-Mar-12 16:08pm

i did it, but it doesn't work with me, in my opinion we should make either number of thread is constant or the dimension of array constant.....
don't know what i should do...

Sergey Alexandrovich Kryukov 6-Mar-12 0:21am

First of all, I would solve the problem using Parallel Library, things like Parallel.ForEach and see.
You should understand that creation of thread is an expensive operation, so it's easy to make worse.
I think that matrix multiplication is a typical case where parallel processing helps, because input matrices stay intact and multiplication of one row by one column is independent.
So, what's the problem?
--SA

Nagy Vilmos 7-Mar-12 6:15am

No work from you === no work from us.

1 solution

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)

Manfred Rudolf Bihy · Answer 1 · 2012-03-07T04:40:00

Work out multiplication compatibility, (m x n) can multiply (n x o) matrices which will yield a (m x o) matrix. That is not really needed here since we are only looking at square matrices which are n x n and thus have the correct dimensions for multiplication.
Choose block size: b
Create a new matrix with the dimensions (m x o).
Traversing the new matrix in a row by row (or column by column) fashion create tuples in batches size b.
Example with b = 2 and m = n = o = 3 and array indices based on one:
1. (1,1) (1,2)
2. (1,3) (2,1)
3. (2,2) (2,3)
4. (3,1) (3,2)
5. (3,3)
Start a thread for each of the batches. Using Thread.Start you will pass the list of tuples to the thread as well as a reference to the the newly created m x o matrix as well as both original matrices in the order left hand side, right hand side.
In the threads work method you'll do the multiplication of row x by column y for each tuple (x, y) of the list you passed to the thread. The resulting scalar is placed into the new matrix at position (x,y).
In your main program wait for all threads to finish (Join). The multiplication result is now in the newly created matrix from step 3.

That is all there is to it really. The multiplication of a row by a column or matrix multiplication[^] is straight forward enough so I won't go into that.

Regards,

Manfred