Add benchmarks (49x improvement with ff)

Tested on a Intel(R) Core(TM) i5-6300U CPU @ 2.40GHz, with 16GB of RAM.

- Old (using nnum-bigint):
hash           time:   [5.9258 ms 5.9407 ms 5.9587 ms]

- New (using ff):
hash           time:   [120.12 us 121.08 us 122.30 us]

Cargo.toml

@@ -12,3 +12,9 @@ readme = "README.md"
 ff = {package="ff_ce" , version="0.11", features = ["derive"]}
 rand = "0.4"
 
+[dev-dependencies]
+criterion = "0.3"
+
+[[bench]]
+name = "bench_poseidon_hash"
+harness = false

README.md

@@ -7,3 +7,8 @@ Compatible with the Poseidon Go implementation done in https://github.com/iden3/
 
 ## Warning
 Do not use in production
+
+## Benchmarks
+```
+hash           time:   [120.12 us 121.08 us 122.30 us]
+```

benches/bench_poseidon_hash.rs

@@ -0,0 +1,30 @@
+use criterion::{black_box, criterion_group, criterion_main, Criterion};
+
+extern crate rand;
+#[macro_use]
+extern crate ff;
+use ff::*;
+
+use poseidon_rs::Poseidon;
+
+fn criterion_benchmark(c: &mut Criterion) {
+    let b1: poseidon_rs::Fr = poseidon_rs::Fr::from_str(
+        "12242166908188651009877250812424843524687801523336557272219921456462821518061",
+    )
+    .unwrap();
+    let b2: poseidon_rs::Fr = poseidon_rs::Fr::from_str(
+        "12242166908188651009877250812424843524687801523336557272219921456462821518061",
+    )
+    .unwrap();
+    let mut big_arr: Vec<poseidon_rs::Fr> = Vec::new();
+    big_arr.push(b1.clone());
+    big_arr.push(b2.clone());
+    let poseidon = Poseidon::new();
+
+    c.bench_function("hash", |b| {
+        b.iter(|| poseidon.hash(big_arr.clone()).unwrap())
+    });
+}
+
+criterion_group!(benches, criterion_benchmark);
+criterion_main!(benches);