Identity pallet benchmark

Added a genesis builder with prepared identities to run the benchmark using --chain=gdev-benchmark.
Added pallet identity benchmark.
Updated the runtime configuration to use the computed weights using a placeholder until runtime/common/src/weights/pallet_identity.rs has been generated using the reference machine.

resolved all threads

marked this merge request as ready

added 1 commit

e3231808 - doc(benchmark): add some air in the benchmark doc

Compare with previous version

I tried

./target/release/duniter benchmark pallet --chain=gdev-benchmark --steps=1 --repeat=1 --pallet=pallet_identity --extrinsic=* --execution=wasm --wasm-execution=compiled --heap-pages=4096 --header=./file_header.txt --output=.

and got

Error: Input("No benchmarks found which match your input.")

Where is the mistake?

Yes, I will commit the Cargo.lock with the "sp-keystore" that is necessary to run the tests of the pallet.

To clarify the point 5 of the documentation:
At the beginning, the pallets/identity/src/weights.rs does not exist at all and the pallet doesn't use any weight info. We run the benchmark to create automatically runtime/common/src/weights/pallet_identity.rs. We use this file to create manually the pallets/identity/src/weights.rs that will contain the trait definition and a dummy implementation for testing purpose. After that, we can use the weight info inside the pallet that will run on test and runtime. At the end, the benchmark can be run again, but on the reference machine to regenerate runtime/common/src/weights/pallet_identity.rs.

added 4 commits

3367f6b9 - feat: add pallet identity benchmark
cf9afc5f - feat(pallet_identity): add weights info
d7577268 - feat(runtimes): use our benchmarks for pallet identity
a8332ffa - doc(benchmark): add some air in the benchmark doc

Compare with previous version

As we will not squash the commits from this MR, I suggest:

detailing point 5 of the doc with what you describe in previous post
1. create runtime/common/src/weights/pallet_<pallet>.rs automatically
2. add pallets/<pallet>/src/weights.rs manually
reflect these steps in the commit history so that the topological order matches the previous chronological order

This way, the commits from this MR could serve as an example of the steps detailed in the doc.

Identity pallet benchmark

Merge request reports

Activity