benchmark the use of different rounds when creating a new ssh key with `ed25519` type in order to balance strength with speed

readme.md

anything in the range close to 1 second is good enough

remeber a strong passphrase is better than a high number of rounds

---

$ lscpu | egrep -i 'thread|core|socket'
Thread(s) per core:                   2
Core(s) per socket:                   6
Socket(s):                            1

---

Benchmarking ssh-keygen with Ed25519 on CPU: AMD Ryzen 5 7530U

Testing with 100 rounds...

Key generation took: .53 seconds

Testing with 200 rounds...

Key generation took: 1.05 seconds

Testing with 500 rounds...

Key generation took: 2.64 seconds

Testing with 1000 rounds...

Key generation took: 5.28 seconds

benchmark-ssh-key-ed25519-rounds.sh

#!/bin/bash

cpu_model=$(lscpu | grep "Model name" | awk -F: '{print $2}')
# cpu_model=$(lscpu | grep "Model name" | awk -F: '{print $2}' | tr -d '[:space:]')

echo "Benchmarking ssh-keygen with Ed25519 on CPU: $cpu_model"

rounds_values=(100 200 500 1000)
# rounds_values=(100 200 500 1000 1500 2000 3000 4000 5000)
for rounds in "${rounds_values[@]}"; do
    echo "Testing with $rounds rounds..."
    start_time=$(date +%s.%N)
    ssh-keygen -t ed25519 -a "$rounds" -f /tmp/test_key -N "this-is-a-long-passphrase-and-vulnerable-one" -q &>/dev/null
    end_time=$(date +%s.%N)
    duration=$(echo "$end_time - $start_time" | bc)
    echo "Key generation took: $duration seconds"
    rm /tmp/test_key # Clean up the temporary key
done