Intel® oneAPI Math Kernel Library (oneMKL) - Data Parallel C++ Developer Reference

Developer Reference

Contents

Random Number Generators Device Routines

The main purpose of Device routines is to make them callable from your DPC++ kernels; however, there are no limitations to be called from the Host. For example:
sycl::queue queue;

queue.submit([&](sycl::handler& cgh) {
  cgh.parallel_for(range,[=](…) {
     mkl::rng::device::routine(...); // calling routine from user’s kernel code
  });
});

mkl::rng::device::routine(...); // calling routine from host
RNG Device APIs Structure
RNG Device API Structure
RNG Device API Structure

Product and Performance Information

1

Performance varies by use, configuration and other factors. Learn more at www.Intel.com/PerformanceIndex.