Tracking: optimization of cryptographic algorithm implementations

[FriendlyCM]

Describe the feature
This tracking issue includes all work related to the optimization of cryptographic algorithm implementations, including related sub-issues and pull requests.

Related PRs

• feat(precompiles): optimize bn128 precompiles #5507
• feat(precompiles): optimize native implementation of the bn128 precompiles #5611

[Federico2014]

About one month ago, we started two full nodes to test the optimized BN128 precompiled contracts and the ECKeyV2 synchronization. The first full node began syncing from block 0, while the second node started from block 51,156,712.

Currently, the first node has synced up to block 34,569,049, and the second node has synced up to 73,063,249. So far, synchronization has been running smoothly on both nodes. During the sync process, we also monitored whether any on-chain contracts were invoking BN128 precompiled contracts. The results show:

• For BN128Addition and BN128Multiplication, they have been called by contract with method signature 73fa58f5. Its source code can be traced to here.
• For the BN128Pairing precompiled contract, we found multiple contract addresses—such as TLsHCcjPY7U6PFrZtJoVEbraKNL5dexyp9, TBd2mjyFMS2BKLZvJ3aJdVn43PBBk45737, and TEFGPUn8zJoHvod7bZbRRytnPmz32QYSLu—invoking it. All of these contracts were deployed by the same address, and none of them has published source code. A small number of BN128Pairing transactions succeeded, for example: tx. However, most transactions failed, such as: tx. Failures appear to be caused by reasons such as transaction timeout or invalid execution.

These findings indicate that although a few contracts attempt to call the BN128Pairing precompile on mainnet, the majority of these transactions fail, meaning the feature is effectively unusable in its current state. Therefore, optimizing the BN128 precompiled contracts is necessary.

[halibobo1205]

Closing this tracking issue. Further progress will be tracked in #6374 — please follow updates there for the latest status. Thanks @Federico2014!