Control access to global data

Question

0.00/5 (No votes)

See more:

I have a code which has a loop which takes around 2 hrs to complete one iteration. I have to execute 12 iterations and therefore the code takes around 24 hrs to complete. I have a 16 core processor to my access. I wish to parallelize the loop such that I have 12 threads, each executing one iteration on one core. This will bring down the execution time to 2 hrs. I have achieved the parallel code but however, there is a conflict for the data to be read from from a global array. How do I resolve that?
My code is in c++

Posted 8-Jun-11 2:59am

neha.arora44c

Add a Solution

Comments

Niklas L 8-Jun-11 9:16am

There can be no conflict reading data. Are you writing to the array as well? It's impossible to answer your question without more information on how you do it.

4 solutions

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)

Espen Harlinn · Answer 1 · 2011-06-08T05:47:00

As Niklas Lindquist mentions: If the data is availalbe in memory before the threads start - possibly using a memory mapping - and the data is not altered by the threads. there is no need to protect the memory using a locking mechanism, just access at will. If each thread writes it's results to a separate memory area you will not need a locking mechanism in place for that either.

Locking is only needed when you want to coordinate updates between threads - even if only one is writing and several others are reading.

Best regards
Espen Harlinn

barneyman · Answer 2 · 2011-06-08T03:11:00

Solution 1

don't use global data; it's an anathema to multi-threaded code ...

By all means use a data pointer which is created by the master thread and used by the children ....

What conflict are you experiencing? reading the data should be fine, however, if you're writing data into anything that's shared you must use some form of synchronisation

Posted 8-Jun-11 3:11am

barneyman

Comments

Sergey Alexandrovich Kryukov 17-Jun-11 3:22am

It's not about "global", it's about "shared". You're basically right about anathema, but in real time not sharing data at all is simply not realistic. My 4.
--SA

Albert Holguin · Answer 3 · 2011-06-08T05:08:00

Solution 2

Use a mutex variable if you need to control access to a data array from multiple threads.

Posted 8-Jun-11 5:08am

Albert Holguin

Comments

Sergey Alexandrovich Kryukov 17-Jun-11 3:24am

I would say "critical section" first (we do not know platform though) if you're talking about threads in the same process. Mutex (named) can also do IPC; same thread synchronization, but working in different processes as well. My 4.
--SA

neha.arora44c · Answer 4 · 2011-06-11T04:09:00

Solution 4

Thanks. :)
Apparently the conflict was in writing the data to the files... All of them were accessing the same file together.

Posted 11-Jun-11 4:09am

neha.arora44c

Control access to global data

4 solutions

Solution 3

Solution 1

Solution 2

Solution 4

Add your solution here

Preview 0

Existing Members

...or Join us