Databases
StackOverflow2013
Postsdbo
PO
 

Note that there are some explanatory texts on larger screens.

plurals
  1. PO
    primarykey
    data
    text
    <p>There are not a lot of "good" things that can come from playing around with thread stack sizes. Usually when this sort of thing comes up, it's because someone is trying to do create hundreds (or thousands) of threads and running into per process limits. This is generally not a good way to use threads. Example: creating a new thread for every incoming connection request to a server.</p> <p>Even if that's not why you are asking, a few things you should know. </p> <p>In production (real world) code, not a simple example, you can't ever be sure that a specific "reduced" stack size per thread is 100% safe. You can test it extensively at a given stack size and then maybe once you're convinced in your own test environment that it's "large enough", probably increase it slightly (maybe 10-15%) for production use. I don't know of any specific tools that can tell you the peak stack usage per thread at run time. Better off to avoid this completely.</p> <p>Reducing the stack size per thread will not reduce memory consumption directly, or improve performance. You're not optimizing your code by experimenting with this.</p> <p>It's much safer to use the default stack size chosen by the local implementation when you can, until you have a real reason to change it.</p> <p>You were missing some header files for prototypes, had missing return values in the thread funcs, etc. plus some other slight cleanups.</p> <p>Slightly modified version of your code that runs here:</p> <pre><code>#include &lt;stdio.h&gt; #include &lt;string.h&gt; #include &lt;pthread.h&gt; #include &lt;stdlib.h&gt; #include &lt;assert.h&gt; #include &lt;unistd.h&gt; #include &lt;limits.h&gt; static void *threadFunc1(void *arg) { char arr[5000]; printf("Hello fromt threadFunc1\n"); return NULL; } static void *threadFunc2(void *arg) { char arr[10000]; printf("Hello fromt threadFunc2\n"); return NULL; } int main(int argc, char *argv[]) { pthread_t t1,t2; pthread_attr_t thread_attr; int s = 0; size_t tmp_size=0; s = pthread_attr_init(&amp;thread_attr); assert(s==0); s = pthread_attr_setstacksize(&amp;thread_attr , PTHREAD_STACK_MIN + 0x1000); assert(s==0); s = pthread_attr_getstacksize(&amp;thread_attr , &amp;tmp_size ); assert(s==0); printf("forced stack size of pthread is:%zu\n", tmp_size); s = pthread_create(&amp;t1, &amp;thread_attr, threadFunc1, NULL); assert(s==0); sleep(1); /* not the "right" way, but ok for a quick test */ s = pthread_create(&amp;t2, &amp;thread_attr, threadFunc2, NULL); assert(s==0); sleep(1); printf("Main done()\n"); return 0; } </code></pre> <p>Note: I used </p> <pre><code>$ gcc -O0 -Wall stack_t.c -o stack_t -pthread ./stack_t </code></pre> <p>If you want to experiment with other sizes for the stack, be aware that some implementations require the stacksize to be a multiple of the system page file size. You can try values larger than PTHREAD_STACK_MIN as well obviously, just be aware of that potential issue. </p> <p>EDIT: The example above adds a value to the minimum as an example of increasing it slightly. You may need to add even more. If 0x1000 is not enough, try 0x2000, 0x3000, etc. until you find one that works. As you have discovered, reducing the thread stack size is tricky and can be platform dependent as well. Which is why I have warnings above about it being unsafe.</p>
    singulars
    1. This table or related slice is empty.
    1. POHow can I decide how much stack I can use after a call to pthread_attr_setstacksize?
      singulars
      1. PTQuestion
    1. PTAnswer
    1. USRandy Howard
    1. USRandy Howard
    plurals
    1. This table or related slice is empty.
    1. This table or related slice is empty.
    1. This table or related slice is empty.
    1. This table or related slice is empty.
    1. This table or related slice is empty.
    1. COI am not intrested in running some specific amount of threads or limit memory usage. I was just curious if I could reduce default stack size of thread to see if my code overrunned it. The code I wanted to run is not run on linux at all in my production environment so that is the reason to provide a simple example as an illustration.
      singulars
      1. PO
      1. USkungjohan
    2. COIf I understand you correctly, you want to reduce the thread stack size, just to find out when it will "break"? That can vary depending upon the code, data sets, etc. About the best you can do is to experiment with reduced stack sizes, and run a lot of test cases with varying data sets, command line arguments, user interactions, etc. (depending upon what the real program does) and find what breaks. This is why the default stack sizes are usually set pretty high in most implementations, to avoid having to go through this process.
      singulars
      1. PO
      1. USRandy Howard
    3. COYes this is for debugging of code that will run on another platform with limited memory. I know exactly what test cases I want to run, but I dont know how to set the exact stack size I want to test.
      singulars
      1. PO
      1. USkungjohan
 

Querying!

 
Guidance
A row detail

Detail views are divided into sections. All the information in the data section comes from columns in the selected row. The other sections display data from other, related rows.

Related data can be related in a to-one or a to-many fashion. Captions of data related in a to-many fashion link to a list view showing a filtered view of the table.

Try moving around until you find a non-empty to-many entry and click on the label to get to one. You can move back to the root by clicking on the database name in the header.

SQuiL has stopped working due to an internal error.

If you are curious you may find further information in the browser console, which is accessible through the devtools (F12).

Reload