Introduce per-device submission threads in the SYCL backend #4670

Job	Run time
read-build-matrix	8s
clang-tidy	1m 43s
build-and-test (dpcpp, ad494e9d, false, 22.04, intel, Debug)	3m 23s
build-and-test (dpcpp, ad494e9d, false, 22.04, intel, Release)	2m 41s
build-and-test (dpcpp, latest, true, 24.04, intel, Debug)	4m 6s
build-and-test (dpcpp, latest, true, 24.04, intel, Release)	2m 59s
build-and-test (acpp, v24.06.0, false, 22.04, nvidia, Debug)	10m 17s
build-and-test (acpp, v24.06.0, false, 22.04, nvidia, Release)	12m 51s
build-and-test (acpp, latest, true, 24.04, nvidia, Debug)	3m 2s
build-and-test (acpp, latest, true, 24.04, nvidia, Release)	2m 6s
build-and-test (simsycl, HEAD, true, 24.04, intel, Release)	1m 34s
build-and-test (simsycl, HEAD, true, 24.04, intel, Debug)	6m 2s
report	21s
tag-latest-containers	0s
	51m 13s

Provide feedback