diff --git a/readme.md b/readme.md
--- a/readme.md
+++ b/readme.md
@@ -1,7 +1,42 @@
 # Shamira #
 
-Implements Shamir's secret sharing algorithm. Splits a string or a byte sequence byte-per-byte into n<255 shares, with any k of them sufficient for reconstruction of the original input.
+Implements Shamir's secret sharing algorithm. Splits a string or a byte sequence byte-per-byte into _n_<255 shares, with any _k_ of them sufficient for reconstruction of the original input.
 
 Outputs the shares as hexadecimal, Base32 or Base64 encoded strings.
 
-Can be used on its own from the command line by invoking shamira.py or as a library by importing shamira.py.
+Can be used on its own from the command line by invoking `shamira.py` or as a library by importing `shamira.py`.
+
+## Performance ##
+
+As it is, the code is not very fast. Splitting takes _n_ evaluations of a polynomial of order _k_ over Galois field 256, leading to _O(n*k)_ finite field multiplications. Reconstruction of the constant parameters during joining similarly takes _O(k*k)_ multiplications.
+
+Benchmark results, all values mean _seconds per byte_ of the secret length:
+<table>
+    <tr>
+        <th>k / n parameters</th>
+        <th>Split</th>
+        <th>Join</th>
+    </tr>
+    <tr>
+        <td>2 / 3 (a Raspberry Pi 3)</td>
+        <td>7.99e-05</td>
+        <td>0.000428</td>
+    </tr>
+    <tr>
+        <td>2 / 3 (a laptop)</td>
+        <td>1e-05</td>
+        <td>6.7e-05</td>
+    </tr>
+    <tr>
+        <td>254 / 254 (a Raspberry Pi 3)</td>
+        <td>0.417</td>
+        <td>0.471</td>
+    </tr>
+    <tr>
+        <td>254 / 254 (a laptop)</td>
+        <td>0.0431</td>
+        <td>0.0542</td>
+    </tr>
+</table>
+
+While the speeds are not awful, for longer secrets I recommend encrypting them with a random key of your choice and splitting only the key. Anyway, you can run your own benchmark with `benchmark.py`