My research focuses on high-performance concurrent data structures and algorithms, programming language runtime (Go and Rust), standard library, and LLM inference optimization.

• mixhash: a simple hash algorithm specifically designed for uint64 hashmap
• Build a High-Performance sorting algorithm in Go[CN]
• Scalable MPMC queue in Go language[CN]
• Hash Algorithms and Pseudorandom Number Generators[CN]

• GopherCon 2024: Building a High-Performance Concurrent Map in Go

PinnedLoading

1. skipsetskipsetPublic

   skipset is a high-performance, scalable concurrent sorted set based on skip-list. Up to 15x faster than sync.Map in the typical pattern.

   Go 255 19

2. skipmapskipmapPublic

   skipmap is a high-performance, scalable concurrent sorted map based on skip-list. Up to 10x faster than sync.Map in the typical pattern.

   Go 206 16

3. umapumapPublic

   A fast, reliable, simple hashmap that only supports the uint64 key/value. Faster than the runtime hashmap in almost all cases.

   Go 23 4

4. pdqsortpdqsortPublic

   Pattern-defeating quicksort in Go with generics(need Go1.18). About 2x ~ 60x faster than the built-in sort package.

   Go 115 7

5. rust-lang/rustrust-lang/rustPublic

   Empowering everyone to build reliable and efficient software.

   Rust 102k 13.2k

6. golang/gogolang/goPublic

   The Go programming language

   Go 127k 17.9k