Benchmark plutus-cbor

[ch1bo]

In course of formulating a CIP for serialiseData we created this benchmark to compare serialization of our basic on-chain encoder for TxOuts (run via Haskell on the host of course) with the more optimized cborg encoder used via serialise package for Data.

Here is the gist of it:

• Ada-only outputs are not bad
• 20 and definitely 100 asset outputs are serialized much less efficiently
• Makes sense, as those lead to much more appendBytestring calls, which is the builtin used at the core of plutus-cbor

To reproduce these results on your machine: https://github.com/input-output-hk/hydra-poc/blob/8d0a20510a1e0fb007044a3c7aa8ecbf45ddd49c/plutus-cbor/README.md#L40-L45

[github-actions]

Unit Test Results

    6 files  ±0    71 suites  ±0   5m 19s ⏱️ +30s
203 tests ±0  201 ✔️ ±0  2 💤 ±0  0 ❌ ±0 

Results for commit 8d0a205. ± Comparison against base commit a4b843a.