vdb

Vector Database Acceleration Project

This repository contains research and implementation details for GPU-accelerated vector database modules.
It focuses on high-dimensional vector clustering, distance computation, and efficient GPU memory management.

---

📘 Project Sections

• CUDA K-Means Clustering
  GPU-parallelized K-Means algorithm optimized for high-dimensional datasets.

• Asynchronous Vector Distance & Memory Load Manager
  Asynchronous CUDA pipeline with dual-stream architecture for large-scale vector distance computation.

---

📅 Funding Information

Project Title: 초거대 AI 모델의 장기 기억 저장을 위한 벡터 DB 개발
Funding Agency: IITP (SW컴퓨팅산업원천기술개발사업)
Grant No.: RS-2024-00454666
Duration: 2024.7 — 2027.12

---