Optimization

[hwwhww]

What was wrong?

• cache for the encode
  • ref: https://github.com/ethereum/pyrlp/blob/7904316f3da420d8dda214dfe8fa168a8e866cfb/rlp/codec.py#L48-L70
• cache for the roots
  • ref: https://github.com/ethereum/pyrlp/blob/f6e66129bda1833fdefbbc4132d13b1ab3d00fb6/rlp/sedes/serializable.py#L248-L254

How can it be fixed?

Any other optimization ideas?

[jannikluhn]

Any other optimization ideas?

It would be a rather big change and I have no idea how much we would gain (if anything), but it would be interesting to look into memoryviews. We do quite a bit of slicing when deserializing nested objects and with bytes this means a lot of unnecessary copying.