sycl: simplify bin_bcast_kernel

[AD2605]

This PR simplifies the bin-bcast kernel, and adds a special code-path when the all the inputs are contiguous, thus avoiding the need for unnecessary index calculations.

The current bin-bcast launches a 3D grid or a 1D grid, and the former often limiting in the number of workitems it can accomodate.

This PR completely flattens the kernel, which also makes it easier to check for contiguous memory accesses, and the separate contiguous path also opens the possibility of vectorization later on, though in my current testing, it did not prove to bring about meaningful difference to performance.

This PR also bring minor but consistent improvement of around 1 tk/s on some models.

Performance compared with the following parameters (with -mmp 0 -ngl 99 -t 8)

Intel Lunar Lake 140V iGPU

Model	This PR (5a0e7a9 ) tk/s	Master (814f795) tk / s
qwen2 1.5B Q4_0	34.05 ± 0.53	33.39 ± 0.52
gemma2 2B Q4_K	25.00 ± 0.25	24.74 ± 0.21
llama 8B Q4_K - Medium	11.10 ± 1.27	10.2 ± 0.70

Intel Data Center Max 1100

Model	This PR (5a0e7a9 ) tk/s	Master (814f795) tk / s
qwen2 1.5B Q4_0	95.06 ± 0.76	92.51 ± 5.88
gemma2 2B Q4_K	84.62 ± 0.18	82.44 ± 0.22
llama 8B Q4_K - Medium	36.72 ± 0.05	36.35 ± 0.07