Systems Engineer | High-Performance Computing | C++ Optimization

Computer Science student at VNU-UET focused on Hardware Sympathy—optimizing memory layouts, minimizing CPU cache misses, and reducing instruction latency. I engineer software that respects the physical constraints of the hardware.

A study in transforming a legacy simulation into a high-throughput engine using Data-Oriented Design and Linux Systems Programming.

Key Engineering Decisions:

• Rendering: Bypassed SDL2 immediate mode by implementing Texture Locking for direct pixel buffer manipulation (CPU $\to$ VRAM).
• Memory: Removed non-deterministic heap allocations (std::list) in the hot path, replacing them with Stack-Allocated Ring Buffers (std::array) to eliminate GC-like pauses.
• Diagnostics: Profiling via perf and Hotspot to identify and eliminate driver overhead.

Engineering Roadmap (Q1-Q2 2026):

• Modernization: Refactoring legacy "C with Classes" patterns to idiomatic C++20 (Concepts, Ranges, Strong Types).
• Concurrency: Decoupling Render and Logic loops into isolated threads using a lock-free SPSC queue, utilizing CPU Affinity (Thread Pinning) to prevent core migration and cache thrashing.
• Networking: Implementing a LAN PvP layer using non-blocking UDP sockets for real-time state synchronization (opponent view).
• SIMD: Exploring AVX2 intrinsics to vectorize collision detection (Stretch Goal).

View Repository

Languages

• C++20: Concepts, Move Semantics, RAII.
• C: Raw memory management, Linux System Calls.
• Python: Systems automation and build scripts.

Systems & Tools

• Environment: Linux (Arch/EndeavourOS), Bash, Vim.
• Build/Debug: CMake (FetchContent), GDB, Linux perf.
• Virtualization: Docker, Linux Namespaces/Seccomp.

Core Competencies

• Data-Oriented Design (DOD)
• Low-Latency Audio/Video Pipeline
• Lock-free Concurrency
• Algorithm Optimization (Graph Theory, Flow Networks)