Start from the skeleton. Look for the lines //TODO
In this exercise you need to implement a basic measurement of the execution time for a kernel performing matrix-matrix multiplication. The solution of this task will serve as starting point for Memory Optimizations II exercise
First modify the queue definition and enable profiling
queue q{property::queue::enable_profiling{}};Next set-up sycl::event object the same way is done in the previous exercise. Compute the execution time of the kernel by taking the difference between the end of the execution of the kernel and the start of the execution.
e.get_profiling_info<info::event_profiling::command_end>() - e.get_profiling_info<info::event_profiling::command_start>();
Before computing the time you will have first to synchronize the host and the device (e.wait())!