refactor: speed up poseidon with scratch reuse and pow5 fast-path

[sergeytimoshin]

• Reuse a per-hasher scratch buffer for MDS to avoid per-round allocations and collect overhead.
• Cache width to reduce iterator cost.
• Add a fast-path pow5 for alpha == 5 (x5 S-box).

Bench:

cargo bench -p light-poseidon --bench bn254_x5 -- --baseline committed 
   Compiling light-poseidon v0.4.0 (/Users/tsv/src/light-poseidon/light-poseidon)
    Finished `bench` profile [optimized] target(s) in 7.01s
     Running benches/bn254_x5.rs (target/release/deps/bn254_x5-b345d9a579f3f719)
Gnuplot not found, using plotters backend
poseidon_bn254_x5_1     time:   [5.3874 µs 5.3947 µs 5.4025 µs]
                        change: [-45.132% -44.990% -44.839%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 3 outliers among 100 measurements (3.00%)
  2 (2.00%) high mild
  1 (1.00%) high severe

poseidon_bn254_x5_2     time:   [8.9322 µs 8.9634 µs 8.9949 µs]
                        change: [-36.247% -36.016% -35.785%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 6 outliers among 100 measurements (6.00%)
  2 (2.00%) low mild
  4 (4.00%) high mild

poseidon_bn254_x5_3     time:   [13.616 µs 13.667 µs 13.723 µs]
                        change: [-27.183% -26.882% -26.552%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 8 outliers among 100 measurements (8.00%)
  2 (2.00%) low mild
  6 (6.00%) high mild

poseidon_bn254_x5_4     time:   [20.578 µs 20.629 µs 20.682 µs]
                        change: [-20.783% -20.378% -19.971%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 4 outliers among 100 measurements (4.00%)
  1 (1.00%) low mild
  2 (2.00%) high mild
  1 (1.00%) high severe

poseidon_bn254_x5_5     time:   [28.326 µs 28.410 µs 28.500 µs]
                        change: [-16.608% -16.272% -15.928%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 2 outliers among 100 measurements (2.00%)
  1 (1.00%) low mild
  1 (1.00%) high mild

poseidon_bn254_x5_6     time:   [38.968 µs 39.091 µs 39.236 µs]
                        change: [-13.448% -13.124% -12.810%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 5 outliers among 100 measurements (5.00%)
  3 (3.00%) low mild
  2 (2.00%) high mild

poseidon_bn254_x5_7     time:   [50.393 µs 50.560 µs 50.754 µs]
                        change: [-10.629% -10.243% -9.8770%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 3 outliers among 100 measurements (3.00%)
  1 (1.00%) low mild
  2 (2.00%) high mild

poseidon_bn254_x5_8     time:   [61.839 µs 61.986 µs 62.139 µs]
                        change: [-9.3874% -9.0479% -8.7104%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 3 outliers among 100 measurements (3.00%)
  2 (2.00%) low mild
  1 (1.00%) high mild

poseidon_bn254_x5_9     time:   [72.613 µs 72.821 µs 73.034 µs]
                        change: [-8.0375% -7.6361% -7.2427%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 1 outliers among 100 measurements (1.00%)
  1 (1.00%) low mild

poseidon_bn254_x5_10    time:   [94.341 µs 94.606 µs 94.877 µs]
                        change: [-6.2655% -5.8588% -5.4217%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 1 outliers among 100 measurements (1.00%)
  1 (1.00%) low mild

poseidon_bn254_x5_11    time:   [102.48 µs 102.72 µs 102.96 µs]
                        change: [-7.4800% -7.0893% -6.6792%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 4 outliers among 100 measurements (4.00%)
  2 (2.00%) high mild
  2 (2.00%) high severe

poseidon_bn254_x5_12    time:   [127.98 µs 128.31 µs 128.65 µs]
                        change: [-6.4339% -6.0649% -5.7056%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 2 outliers among 100 measurements (2.00%)
  1 (1.00%) low mild
  1 (1.00%) high mild