I implemented a solution that would produce Everything looks good! ヽ(‘ー`)ノ
message, and in results.edn I have following stats:
- median latency 626
- max latency 1052
- messages-per-op 28
I achieved this with following techniques:
- node uses suggested topology
- node keeps track of messages that its neighbours have received
- every 100ms node sends delta to every neighbour
- node have a mutex similar to maelstrom one to keep gossip task in sync with everything
My question is does maelstrom test results depend on machine specs? If they dont, where I can look for tips? Also why in challenge #3e it should be less messages per operation, but more latency? Am I undestanding correctly that challenges designed to optimize different metrics?