[BUG] The number of "host_cores_used" is incorrect on Intel EMR

[LizYou]

Describe the bug
The "host_cores_used" report in cuvs benchmark output is incorrect on Intel EMR.

Steps/Code to reproduce bug
./FAISS_CPU_IVF_PQ_ANN_BENCH --build --data_prefix=./datasets --benchmark_out_format=json --benchmark_counters_tabular=true --benchmark_ouq,large,k10,bs8_5fd8e2d4-c991-11ef-b9f3-be3af2b6059f.json
dataset: deep-image-96-inner
dim: 96
distance: euclidean
host_cores_used: 48 -----------> this is the core numbers per socket not the total physical cores.
host_cpu_freq_max: 4000000000
host_cpu_freq_min: 800000000
host_pagesize: 4096
host_processors_sysconf: 192
host_processors_used: 192
host_total_ram_size: 540659675136
host_total_swap_size: 8589930496
n_records: 9990000

Expected behavior
The "host_cores_used" should consider the cores across socket.
Suggested Fix: in file ./cpp/bench/ann/src/common/util.hpp for calculating the host_cores_selected, should consider different "physical id" .

Environment details (please complete the following information):

• Environment location: [Bare-metal, Docker, Cloud(specify cloud provider)]
• Method of RAFT install: [conda, Docker, or from source]
  • If method of install is [Docker], provide docker pull & docker run commands used

Additional context
Add any other context about the problem here.