How do we know that the code is thread safe in sample application?

How do we know that the code is thread safe in sample application?

I am following this tutorial  to understand how to remove locks , using VTUNE  
This page says the following after collecting Vtune results:  

**Identify the Hottest Code Lines**
Click the  hotspot navigation button
 to go to the code line that took the most Wait time. VTune Amplifier
 highlights line 170 entering the critical section rgb_critical_section
 in the draw_task function. The draw_task function was waiting for
 almost 27 seconds while this code line was executing and most of the
 time the processor was underutilized. During this time, the critical
 section was contended 438 times.
 The rgb_critical section is the place where the application is
 serializing. Each thread has to wait for the critical section to be
 available before it can proceed. Only one thread can be in the
 critical section at a time. You need to optimize the code to make it
 more concurrent.

I was able to follow this tutorial until I reach the next section: Remove the Lock 

**Remove the Lock**

> The rgb_critical_section was introduced to protect calculation from
> multithreaded access. The brief analysis shows that the code is thread
> safe and the critical section is not really needed.

My question is how do we know the code is thread safe? 

As suggested I commented those line (EnterCritical... and LeaveCritical...), and saw huge performance gain, but I did not  get why this critical section is not required? Which analysis tells us this ?  

The relevant code is here in analyze_locks.cpp: 

        void operator () (const tbb::blocked_range <int> &r) const {
            unsigned int serial = 1;
            unsigned int mboxsize = sizeof(unsigned int)*(max_objectid() + 20);
            unsigned int * local_mbox = (unsigned int *) alloca(mboxsize);
            for (int y=r.begin(); y!=r.end(); ++y) {
                drawing_area drawing(startx, totaly-y, stopx-startx, 1);
                // Enter Critical Section to protect pixel calculation from multithreaded access (Needed?)
            //    EnterCriticalSection(&rgb_critical_section);
                for (int x = startx; x < stopx; x++) {
                    color_t c = render_one_pixel (x, y, local_mbox, serial, startx, stopx, starty, stopy);
                // Exit from the critical section
            //    LeaveCriticalSection(&rgb_critical_section);
                if(!video->next_frame()) tbb::task::self().cancel_group_execution();
        draw_task () {}


2 posts / 0 new
Last post
For more complete information about compiler optimizations, see our Optimization Notice.

There is no conflict in parallelsim, because all pixels' position (x,y) are different.

Critcal Section here is unnecessary, you can use another tool Inspector XE to do thread check.

Leave a Comment

Please sign in to add a comment. Not a member? Join today