ddsha441981

ddsha441981/pulse_map

A CPU cache-line hash table with zero-cost eviction. 64-byte buckets, LFU+LRU hybrid, 16-shard concurrency, per-entry TTL. 6.5-12x faster than moka on 4-thread workloads.

1 stars 1 forks 1 watchers Rust Other

cache cache-line concurrent eviction hashmap lfu lru no-std performance rust sharded ttl

View on GitHub Website

5 Open Issues Need Help Last updated: Aug 17, 2026

Open Issues Need Help

View All on GitHub

Add CI job to verify no_std compilation on thumbv7m-none-eabi about 1 hour ago

help wanted good first issue

ddsha441981/pulse_map

1

A CPU cache-line hash table with zero-cost eviction. 64-byte buckets, LFU+LRU hybrid, 16-shard concurrency, per-entry TTL. 6.5-12x faster than moka on 4-thread workloads.

Rust

#cache#cache-line#concurrent#eviction#hashmap#lfu#lru#no-std#performance#rust#sharded#ttl

Run Miri on test suite to detect undefined behavior in unsafe code about 3 hours ago

help wanted

ddsha441981/pulse_map

1

A CPU cache-line hash table with zero-cost eviction. 64-byte buckets, LFU+LRU hybrid, 16-shard concurrency, per-entry TTL. 6.5-12x faster than moka on 4-thread workloads.

Rust

#cache#cache-line#concurrent#eviction#hashmap#lfu#lru#no-std#performance#rust#sharded#ttl

Add loom tests for MetaWord AtomicU64 CAS correctness about 3 hours ago

help wanted

ddsha441981/pulse_map

1

A CPU cache-line hash table with zero-cost eviction. 64-byte buckets, LFU+LRU hybrid, 16-shard concurrency, per-entry TTL. 6.5-12x faster than moka on 4-thread workloads.

Rust

#cache#cache-line#concurrent#eviction#hashmap#lfu#lru#no-std#performance#rust#sharded#ttl

Add 24-hour soak test for ConcurrentPulseMap and ShardedPulseMap about 3 hours ago

help wanted good first issue

ddsha441981/pulse_map

1

A CPU cache-line hash table with zero-cost eviction. 64-byte buckets, LFU+LRU hybrid, 16-shard concurrency, per-entry TTL. 6.5-12x faster than moka on 4-thread workloads.

Rust

#cache#cache-line#concurrent#eviction#hashmap#lfu#lru#no-std#performance#rust#sharded#ttl

Add cargo-fuzz harness for insert/get/remove sequences about 3 hours ago

help wanted good first issue

ddsha441981/pulse_map

1

A CPU cache-line hash table with zero-cost eviction. 64-byte buckets, LFU+LRU hybrid, 16-shard concurrency, per-entry TTL. 6.5-12x faster than moka on 4-thread workloads.

Rust

#cache#cache-line#concurrent#eviction#hashmap#lfu#lru#no-std#performance#rust#sharded#ttl