Skip to content

Chore/first benchmark weights#92

Merged
nol4lej merged 7 commits into
mainfrom
chore/first-benchmark-weights
May 23, 2026
Merged

Chore/first benchmark weights#92
nol4lej merged 7 commits into
mainfrom
chore/first-benchmark-weights

Conversation

@nol4lej
Copy link
Copy Markdown
Member

@nol4lej nol4lej commented May 22, 2026

No description provided.

nol4lej added 7 commits May 22, 2026 22:19
Measured on AMD EPYC 8 vCPU / 32 GB RAM (reference hardware).

pallet_account_mapping  – 16 extrinsics refreshed
pallet_shielded_pool    – shield, shield_batch, private_transfer,
                          unshield, register_asset, verify_asset,
                          unverify_asset
pallet_zk_verifier      – verify_proof now reflects real Groth16
                          pairing cost (~12 250 µs on BN254)
pallet_relayer          – all extrinsics refreshed
… weights

pallet-account-mapping  0.2.0 → 0.2.1  (CHANGELOG created)
pallet-relayer          0.2.0 → 0.2.1
pallet-shielded-pool    0.8.0 → 0.8.1
pallet-zk-verifier      0.7.0 → 0.7.1

All CHANGELOGs document the Hetzner CCX33 weight measurements
(AMD EPYC-Milan, 8 vCPU, 32 GB RAM, steps=50, repeat=20).
The Makefile test-release target runs tests twice, the second pass with
--features=runtime-benchmarks. In that pass the #[cfg] block in
claim_shielded skips ZK verification, so tests expecting Err(InvalidProof)
returned Ok(()) instead.

Guard the two rejection tests and their helper make_rejected_proof with
#[cfg(not(feature = "runtime-benchmarks"))] so they are excluded when
the VALUE_PROOF circuit is intentionally not registered in the benchmark
runtime.
@nol4lej nol4lej merged commit 9cf6066 into main May 23, 2026
6 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant