Add ASV benchmark suite for dpctl

[vchamarthi]

Adds benchmarks/ covering: object construction, get_device_cached_queue, USM alloc/free, device enumeration, kernel bundle compile (SPIR-V/OpenCL/SYCL source, cold vs. warm), data movement, kernel submit.

Coverage (53 benchmarks, 7 files)

File	Class	API exercised
bench_construct.py	Construct	SyclDevice/SyclContext/SyclQueue construction, from selector, from device, from (context, device), in_order property
	ConstructPlatform	SyclPlatform() construction
	DeviceProperties	name/driver_version/max_compute_units/max_work_group_size/global_mem_size/has_aspect_fp64
	QueueAccessors	sycl_device/sycl_context accessors
bench_queue_cache.py	QueueCache	get_device_cached_queue (context, device) key, device key, filter-string key, vs. an uncached baseline
bench_usm.py	USMAllocation	MemoryUSM{Device,Host,Shared} alloc/free, aligned alloc (alignment=4096)
	USMFirstTouch	alloc + first write + free
	USMQueries	get_usm_type, __sycl_usm_array_interface__
bench_enumerate.py	Enumerate	get_devices/get_num_devices × device_type × backend
	Select	select_default_device/select_cpu_device/select_gpu_device/has_cpu_devices/has_gpu_devices/select_device_with_aspects
bench_compile.py	BundleFromSPIRV	create_kernel_bundle_from_spirv, get_sycl_kernel/has_sycl_kernel
	BundleFromOpenCLSource	create_kernel_bundle_from_source cold (unique kernel name) vs. warm (cache hit)
	BundleFromSYCLSource	create_kernel_bundle_from_sycl_source cold vs. warm
	SourceCompilationProbe	SyclDevice.can_compile("sycl")
bench_copy.py	HostDeviceCopy	SyclQueue.memcpy/memcpy_async, _Memory.copy_from_host/copy_to_host
	DeviceDeviceCopy	SyclQueue.memcpy (d2d), _Memory.copy_from_device
	Fill	SyclQueue.memset/fill/memset_async, _Memory.memset
bench_submit.py	Submit	submit, submit_async (single + batched)
	Synchronize	submit_barrier, SyclQueue.wait, SyclEvent.wait

• Have you provided a meaningful PR description?
• Have you added a test, reproducer or referred to an issue with a reproducer?
• Have you tested your changes locally for CPU and GPU devices?
• Have you made sure that new changes do not introduce compiler warnings?
• Have you checked performance impact of proposed changes?
• Have you added documentation for your changes, if necessary?
• Have you added your changes to the changelog?
• If this PR is a work in progress, are you opening the PR as a draft?

[coveralls]

coverage: 74.959%. remained the same — vchamarthi:asv-benchmarks into IntelPython:master

[ndgrigorian]

I think all of these comments can be cleaned up to be less verbose/unnecessarily specific (i.e., we don't need the phrase 'USM kinds, mapped to their constructors', we can just say 'USM kinds')