setrlimit() not affecting spawned std::threads

Question

I am currently working on a pipeline which loads and transforms multiple images at once. As this is happening to many images at the same time (1440) the memory footprint is quite heavy. I therefore tried to implement a memory management system based on setrlimit, however it doesn't seem to affect the spawned threads (std::thread) as they will happily ignore the limit - I know this because of calls to getrlimit() in the threaded functions - and eventually cause my program to be killed. Here is the code I use for setting the limit:

void setMemoryLimit(std::uint64_t bytes)
{
    struct rlimit limit;
    getrlimit(RLIMIT_AS, &limit);

    if(bytes <= limit.rlim_max)
    {
        limit.rlim_cur = bytes;
        std::cout << "New memory limit: " << limit.rlim_cur << " bytes" << std::endl;
    }
    else
    {
        limit.rlim_cur = limit.rlim_max;
        std::cout << "WARNING: Memory limit couldn't be set to " << bytes << " bytes" << std::endl;
        std::cout << "New memory limit: " << limit.rlim_cur << " bytes" << std::endl;
    }

    if(setrlimit(RLIMIT_AS, &limit) != 0)
        std::perror("WARNING: memory limit couldn't be set:");

    // included for debugging purposes
    struct rlimit tmp;
    getrlimit(RLIMIT_AS, &tmp);
    std::cout << "Tmp limit: " << tmp.rlim_cur << " bytes" << std::endl; // prints the correct limit
}

I'm using Linux. The man page states that setrlimit affects the whole process so I'm kind of clueless why the threads don't seem to be affected.

Edit: By the way, the function above is called at the very beginning of main().

Answer 1

The problem was quite hard to find as it consisted of two entirely independent components:

1. My executable was compiled with -fomit-frame-pointer. This will result in a reset of the limit. See the following example:

   /* rlimit.cpp */
   #include <iostream>
   #include <thread>
   #include <vector>
   
   #include <sys/resource.h>
   
   class A
   {
       public:
           void foo()
           {
               struct rlimit limit;
               getrlimit(RLIMIT_AS, &limit);
               std::cout << "Limit: " << limit.rlim_cur << std::endl;
           }
   };
   
   int main()
   {
       struct rlimit limit;
       limit.rlim_cur = 500 * 1024 * 1024;
       setrlimit(RLIMIT_AS, &limit);
       std::cout << "Limit: " << limit.rlim_cur << std::endl;
   
       std::vector<std::thread> t;
   
       for(int i = 0; i < 5; i++)
       {
           A a;
           t.push_back(std::thread(&A::foo, &a));
       }
   
       for(auto thread : t)
           thread.join();
   
       return 0;
   }
   

   Outputs:

   > g++ -std=c++11 -pthread -fomit-frame-pointer rlimit.cpp -o limit
   > ./limit
   Limit: 524288000
   Limit: 18446744073709551615
   Limit: 18446744073709551615
   Limit: 18446744073709551615
   Limit: 18446744073709551615
   Limit: 18446744073709551615
   
   > g++ -std=c++11 -pthread rlimit.cpp -o limit
   > ./limit
   Limit: 524288000
   Limit: 524288000
   Limit: 524288000
   Limit: 524288000
   Limit: 524288000
   Limit: 524288000
   

2. For the image processing part I work with OpenCL. Apparently NVIDIA's implementation calls setrlimit and pushes the limit to rlim_max.