paradigmxyz
reth
Blog
Docs
Changelog
Blog
Docs
Changelog
Overview
Branches
Benchmarks
Runs
Performance History
Latest Results
fix(trie): return pruned_roots from prune_by_subtrie for revealed_account_paths cleanup Amp-Thread-ID: https://ampcode.com/threads/T-019c261b-0754-7023-9ec4-b9907a604f14 Co-authored-by: Amp <amp@ampcode.com>
feat/hot-account-tracking
5 minutes ago
ci: enable debug assertions for statetests (#21775)
main
15 minutes ago
feat(metrics): add state_subtries_cleared and state_prune_count metrics Amp-Thread-ID: https://ampcode.com/threads/T-019c261b-0754-7023-9ec4-b9907a604f14 Co-authored-by: Amp <amp@ampcode.com>
feat/hot-account-tracking
15 minutes ago
perf(trie): parallelize subtrie clearing in prune_by_subtrie Co-authored-by: Amp <amp@ampcode.com> Amp-Thread-ID: https://ampcode.com/threads/T-019c261b-0754-7023-9ec4-b9907a604f14
feat/hot-account-tracking
31 minutes ago
perf(trie): parallelize subtrie clearing in prune_by_subtrie Amp-Thread-ID: https://ampcode.com/threads/T-019c261b-0754-7023-9ec4-b9907a604f14 Co-authored-by: Amp <amp@ampcode.com>
feat/hot-account-tracking
32 minutes ago
perf(trie): use fast subtrie-based pruning for state trie Integrate prune_by_subtrie into prune_preserving for state tries. When excess_memory is finite, try the fast O(256) subtrie clearing approach first. Only fall back to full DFS if that's not enough. This should significantly reduce state_prune_duration for typical memory pressure scenarios. Co-authored-by: Amp <amp@ampcode.com> Amp-Thread-ID: https://ampcode.com/threads/T-019c25d9-89b1-711c-a148-70b0233f98de
feat/hot-account-tracking
37 minutes ago
perf(trie): use fast subtrie-based pruning for state trie Integrate prune_by_subtrie into prune_preserving for state tries. When excess_memory is finite, try the fast O(256) subtrie clearing approach first. Only fall back to full DFS if that's not enough. This should significantly reduce state_prune_duration for typical memory pressure scenarios. Amp-Thread-ID: https://ampcode.com/threads/T-019c25d9-89b1-711c-a148-70b0233f98de Co-authored-by: Amp <amp@ampcode.com>
feat/hot-account-tracking
38 minutes ago
feat(trie): add prune_by_subtrie for fast memory-based pruning Add an alternative pruning method that clears entire lower subtries instead of doing depth-based DFS traversal. This is O(256) instead of O(millions) when we only need to free some memory. Algorithm: 1. Compute memory for each of 256 lower subtries (parallel) 2. Sort by size (largest first), excluding hot account subtries 3. Clear subtries until we've freed enough memory This is faster than prune_preserving when the goal is just to reduce memory usage rather than precise depth-based pruning. Co-authored-by: Amp <amp@ampcode.com> Amp-Thread-ID: https://ampcode.com/threads/T-019c25d9-89b1-711c-a148-70b0233f98de
feat/hot-account-tracking
39 minutes ago
Active Branches
feat(trie): add tiered hot account tracking for smart trie preservation
last run
5 minutes ago
#21651
CodSpeed Performance Gauge
0%
chore: improve some spans
last run
51 minutes ago
#21781
CodSpeed Performance Gauge
-11%
chore: better default filter for profiling
last run
51 minutes ago
#21779
CodSpeed Performance Gauge
0%
© 2026 CodSpeed Technology
Home
Terms
Privacy
Docs