Skip to content
Snippets Groups Projects

Solve #426

Merged Daniel Hugo Campora Perez requested to merge dcampora_small_improvements into master
All threads resolved!
2 files
+ 7
1
Compare changes
  • Side-by-side
  • Inline
Files
2
@@ -99,6 +99,12 @@ extern thread_local BlockIndices blockIdx;
constexpr BlockDimensions blockDim {1, 1, 1};
constexpr ThreadIndices threadIdx {0, 0, 0};
/**
* @brief Emulates atomic operations so they work on the CPU backend.
* @details The CPU backend executes sequentially, so there
* is no need to provide atomicity in any of the
* implementations below.
*/
template<class T, class S>
T atomicAdd(T* address, S val)
{
Loading