Created
May 14, 2020 14:57
-
-
Save mratsim/6e7d4a05c526cdfe64c588bdcab860e1 to your computer and use it in GitHub Desktop.
2020-05-14 Eth2 impl call 39
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
https://github.com/ethereum/eth2.0-pm/issues/149 | |
-------------------- | |
Nimbus: | |
- Multiple Sync fixes | |
- Snappy (add compute issue with Lighthouse’s) | |
- can now sync to Schlesi: slowly but steadily | |
- working on performance | |
- in particular Windows | |
- Fixed multiple memory leaks | |
- libp2p | |
- our block caching system | |
- added more memory tracking tools | |
- will add a status button for our testnets | |
- still have an “old libp2p connection” leak | |
Tools: | |
- debug discovery | |
- debug libp2p topics and message received | |
———————— | |
BLS change coming soon: https://github.com/ethereum/eth2.0-pm/issues/149 | |
Networking updates | |
Test vectors updates to catch new corner cases | |
——————————— | |
BeaconFuzz: | |
- Found upstream bug in Snappy | |
- Found infinite loop in Teku | |
- Stack overflow in Nimbus. | |
- Refactoring into 3 different tools | |
Q: on BeaconState being a trusted input | |
—> Overflow in state_transition == invalid transition | |
—> One of the reason to split up BeaocnFuzz | |
https://blog.sigmaprime.io/beacon-fuzz-04.html | |
——————————— | |
Teku | |
- Builtin support for Schlesi | |
- 900MB memory footprint | |
Lodestar | |
- Can start syncing Schlesi, 2k epochs | |
Nimbus | |
Trinity | |
- Continuing port to (?) async framework | |
- Updated to latest BLS API | |
Nethermind | |
- | |
Prysmatic | |
- Topaz | |
- Fixing UX bugs | |
- Fixing network bugs from multi client testate | |
- Aligned to 0.11.2 | |
- Working on 0.12 | |
- Optimizing syncing (100 blocks/s, not taking into account signing) | |
- Slashing service working (1 client with it enabled earned a lot more than the others) | |
Lighthouse | |
- Key derivation | |
- Trail of bits audit: concurrency, slashing | |
- … | |
- Great improvement on memory usage | |
- 300MB for beacon node + 2k validators | |
- Consensus bug | |
- Remove parallelisation for state transition code, not needed with batched BLS verif | |
- Backed the genesis state into the repo | |
- updated to stable future | |
- Update done —> ready for Trail of bits | |
- RPC handling | |
——————————— | |
Testnets | |
Afri: Failures mostly due to network fragmentation | |
- Beacon node rejecting each other for various reason | |
- Different genesis time between Prysmatic and Lighthouse | |
- Somehow managed to launched a multi client testnet 2 weeks ago | |
- Finality was horrible at start due to crashing | |
- Now almost perfect finality for more than a week | |
==> will start working on a new testnet, maybe for June | |
Proto: | |
- Lodestar 1k epoch 10% | |
- Nimbus: | |
- close to the head, 100 blocks distance but then the sync mode changes and stability issue | |
- added on eth2stats | |
==> Danny: coordinated start in June | |
——————————— | |
Research | |
Eth1x-64 write up on ethresearch and repo in ewasm. | |
- https://ethresear.ch/t/the-eth1x64-experiment/7195 | |
- https://ethresear.ch/t/eth1x64-variant-1-apostille/7365 | |
Radical EVM changes will cause issue with existing tooling so might as well drop and chang to EWASM | |
Vitalik: | |
Homomorphic encryption for private data | |
Call to cryptographer for polynomial commitment work | |
Proof-of-custody, see Dankrad post on ethresear.ch | |
TXRX: | |
Mikhail Eth1/Eth2: https://ethresear.ch/t/the-scope-of-eth1-eth2-merger/7362 | |
Fork choice test: | |
- found a bug in Teku | |
- based on a “transpiler” | |
——————————— | |
Networking | |
Discv5 Felix: | |
- Working to fix a spec bug (message that seemingly has a wrong encoding) | |
- Discovery update: how to deal with live network | |
- > include with BLS spec update | |
——————————— | |
Spec update: | |
Has seen bug reports on phase 1, but not really tested. | |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment