Demonstrate Democratizing Power of SYCL with live demo of multi-target code.

[fraggamuffin]

As I am reviewing the slides this weekend, and they look great, the integration from other teams really added value and new perspectives. One thought occurred to me.

Is there any exercise/demo that demonstrates our principle of write once, but run everywhere on different hardware. I know we can say it, but if we can show it with a live demo on Intel, AMD, Nvidia initially and maybe in future extend it to FPGA, AI processors, that would be even more powerful.

We might not do it in time for CPPCON 2021, but for future.

This tutorial should be maintained for the future.

[AerialMantis]

Exercise 13 gives an example of doing load balancing between two devices such as CPU and GPU, though it's written fairly generically so that it still works even if you don't have multiple different devices.