Create Blog Post, with benchmark results like the one made for artano.

mesudip commented 3 months ago

Using the benchmark and test results for v2 and v3 contracts with/without reference script, Write a blogpost showing the progresses made in PlutusV3 in context of a marketplace contract.

Acceptance Criteria

[x] Blogpost content is created.
[x] Blogpost is published in ~cardano-essentials~ Github publicly

mesudip commented 2 months ago

https://github.com/dQuadrant/cardano-marketplace/tree/main/reports/test-sanchonet-cardano-api-8.46

Based on the tests run on the marketplace with latest following was observed:

Script size of PlutusV3 is larger than the PlutusV2 script
Execution unit consumption of PlutusV3 was lower but the tx-fee was higher than PlutusV2
Running multiple parallel marketplace actions with connection to same node socket exhausts the total no of connection. In our example It was reproduced when running simulation with 25 individual flows.

Analysis:

Native PlutusV3 is worse than the PlutusV2 in terms of size and cost. It is possible to have a workaround by not decoding the ScriptContext and directly working on the BuiltinData, but it makes code look messy and a lot of Plutus helper functions cannot be used when writing contract in that way. (Plutarch with PlutusV3 might be a better option)
In order to orchestrate the simulation, it appears it's better to run multiple nodes and multiplex the connection.

mesudip commented 2 months ago

Native PlutusV3 is worse than the PlutusV2 in terms of size and cost. It is possible to have a workaround by not decoding the ScriptContext and directly working on the BuiltinData, but it makes code look messy and a lot of Plutus helper functions cannot be used when writing contract in that way. (Plutarch with PlutusV3 might be a better option)

Martin response: By "directly working on the BuiltinData" you mean something like https://github.com/IntersectMBO/plutus/blob/b989f60c3b59d844bbaca1b9c39cbe89d2087c72/plutus-ledger-api/test-plugin/Spec/ScriptSize.hs#L98, right?

reeshavacharya commented 2 months ago

https://github.com/cardanoapi/chang-hardfork-testing/blob/b378b7c49992d9818eecebc00a97526e12de10c3/changHardForkTesting/app/V3/Spend/TxInfoFields.hs#L64

spannercode commented 3 weeks ago

This is the final document that will be shared with the community from Intersect side:

https://github.com/cardanoapi/hardfork-testing/blob/main/migration.md

cardanoapi / hardfork-testing

Create Blog Post, with benchmark results like the one made for artano. #34

Acceptance Criteria