[SYCL] Adopt the experimental free function extension

[fwyzard]

Access the current nd_item via a free function instead of storing it as a data member of the various SYCL accelerator base classes.

[bernhardmgruber]

LGTM. Here is a further suggestion:

[fwyzard]

I will keep this as a draft while we check the impact on the performance.

[AuroraPerego]

The performance does not change:

nThreads	base@c9dbac0	base + PR
1	148.44 +/- 3.48)	148.32 +/- 0.35)
2	172.73 +/- 0.94)	171.72 +/- 0.60)
4	181.79 +/- 0.16)	182.07 +/- 0.19)
8	182.89 +/- 0.06)	182.94 +/- 0.15)

However, since this feature is still experimental and there is another one similar to it as proposed, maybe it's better to wait a bit and see if the extension gets modified.

The tests done with pixeltrack-standalone@dcf2898.
The numbers are the throughput, i.e. the number of concurrent events processed per second, computed processing 10000 events and nThreads is the number of CPU threads used.

[j-stephan]

LGTM. Feel free to convert this into a real PR.