A hardware-aware guide to data structures for system software engineers.

VS Code extension: Go struct layout, padding, reorder

⚡ Cut 60-80% token cost for DeepSeek V4 Pro on Codex CLI. token-saver + memory skills with prefix-cache optimization.

High-performance limit order book engine with C++ core and Python SDK. Processes 20M+ msgs/sec with µs latency. Supports real crypto/equity data replay, spread/imbalance/impact analytics, and backtesting of VWAP, TWAP, POV, and market-making strategies with reproducible PnL and risk metrics.

Cache & In-Memory optimizations for Rust, revived from the slabs of Sumer.

Learn matrix-matrix multiplication optimization in C++ through progressively optimized DGEMM implementations, from naïve algorithms to AVX-512 and blocked kernels.

CCProf: Lightweight Detection of Cache Conflicts

Freeze Claude Code's prompt prefix so DeepSeek's automatic cache always hits — alignment proxy + coalescing + keepalive, installable as a CC plugin. Measured 64% cheaper on real Claude Code traffic.

Contains implementations of cache-optimized and external memory algorithms.

A cache-optimized AI memory and context management framework for long-running LLM agents, featuring tiered context storage, token budgeting, memory compaction, sub-agent delegation, and structured web intelligence.

Field level cache optimizations for Rust (no_std)

可运行的 C++20 性能工程指南：示例、基准测试与 VitePress 文档站

A dotnet tool for moving project files into the directories specified by the solution (.sln) file.

Cache-aware orchestration for LLM agents. Fork helpers that share cached prefixes, detect cache breaks, and cut token costs by 38%+.

GitHub Action & CLI to analyze binary memory layouts: detect padding, compare diffs, enforce budgets. Parses DWARF debug info for C/C++/Rust/Go.

DeepSeek缓存优化器 v1.1 — Reasonix四支柱 + 语义压缩 (命中率+30%)

AI-powered C code optimizer demonstrating 2.3× performance gains through SIMD vectorization, cache blocking, and algorithmic improvements. Proof that AI optimizes better than humans.

Arhitectura Sistemelor de Calcul - UPB 2020

Modular Spectrum of Pi: reference implementation of the Stride-6 engine. Unifies Chudnovsky's series with DSP polyphase decomposition in Z/6Z. Validated at 100M digits with 95% parallel efficiency. Features a Shared-Nothing architecture to bypass the memory wall and maximize cache alignment.

Comparison of parallel matrix multiplication methods using OpenMP, focusing on cache efficiency, runtime, and performance analysis with Intel VTune.