improve FEC bench

we want to compare

• naive recoding: k-decoding followed by 1-encoding
• komodo recoding: k-recoding with
  k = \#\text{shards}

results

💡 Note
we see that the naive recoding is around 100 times slower compared to the komodo recoding 🎉

💡 Note
the format of the labels is always {curve} / {k}

changed the description

merged

mentioned in commit 70941cae