profile block processor to improve sync performance

[TalDerei]

profile the block processor and sync mechanism in the web, tracked by https://docs.google.com/document/d/1OCVsyi4qn5DlJmdoY9x2mk0dMgcLF577hlC0TD_fpjo/edit

[TalDerei]

expanding benchmarking scope to profile the cost of poesidon hashing for TCT related operations.

[TalDerei]

benchmarking profile for syncing 800k blocks:

• baseline: 280s
• skip trial decryption: 200s (~25% speedup)
• skip trial decryption and poseidon hashing: 100s (~75% speedup)
• streaming 800k blocks in async block querier: depends on internet connection (experiencing average 30s+ on hotspot)
• binary serialization and wasm decoding: ~5s

as the chain state grows, there’s a strong indication that the bottleneck will increasingly converge towards bandwidth (downloading every compact block) rather than compute for fresh wallets; imagining what it will be a year from now.

[TalDerei]

[TalDerei]

minimizing storage accesses and disk i/o will be profiled and investigated as other potential low-hanging fruit – related to #1814.

update: ~8% improvement.