[Openmp-dev] Use of llvm_omp_target_alloc_shared API

[Itaru Kitayama via Openmp-dev]

1 dimensional case seems to be working.

On Sat, Aug 7, 2021 at 3:31 PM Itaru Kitayama <itaru.kitayama at gmail.com> wrote:
>
> Hi,
> Is this supposed to work? It does not on A100.
>
> #include <cstdlib>
> #include <new>
> #include <limits>
> #include <iostream>
> #include <vector>
> #include <omp.h>
>
> extern "C" {
>   void* llvm_omp_target_alloc_shared(size_t, int);
> }
>
> template <class T>
> struct Mallocator
> {
>   typedef T value_type;
>
>   Mallocator () = default;
>   template <class U> constexpr Mallocator (const Mallocator <U>&) noexcept {}
>
>   [[nodiscard]] T* allocate(std::size_t n) {
>     if (n > std::numeric_limits<std::size_t>::max() / sizeof(T))
>       throw std::bad_array_new_length();
>
>     //if (auto p = static_cast<T*>(std::malloc(n*sizeof(T)))) {
>     if (T* p = static_cast<T*>(llvm_omp_target_alloc_shared(n*sizeof(T), 0))) {
>       report(p, n);
>       return p;
>     }
>
>     throw std::bad_alloc();
>   }
>
>   void deallocate(T* p, std::size_t n) noexcept {
>     report(p, n, 0);
>     std::free(p);
>   }
>
> private:
>   void report(T* p, std::size_t n, bool alloc = true) const {
>     std::cout << (alloc ? "Alloc: " : "Dealloc: ") << sizeof(T)*n
>       << " bytes at " << std::hex << std::showbase
>       << reinterpret_cast<void*>(p) << std::dec << '\n';
>   }
> };
>
> template <class T, class U>
> bool operator==(const Mallocator <T>&, const Mallocator <U>&) { return true; }
> template <class T, class U>
> bool operator!=(const Mallocator <T>&, const Mallocator <U>&) { return false; }
>
> int main()
> {
>   std::vector<std::vector<int, Mallocator<int> >,
> Mallocator<std::vector<int, Mallocator<int> > > > v1;
>   v1.reserve(1024);
>
>   auto p = v1.data();
>   v1[0][0] = 999;
>   v1[1][0] = 1;
> #pragma omp target parallel for device(0) is_device_ptr(p)
>   for (int i=0;i < 2;i++) {
>     printf("v1 %d %d\n", i, p[0][0]);
>   }
>
> }