Benchmarking Language Model Performance on 5th Gen Xeon at GCP