AI 연산을 DRAM 안에서 - GDDR6-AiM 기술 소개
Dive into SK하이닉스의 혁신적인 GDDR6-AiM 기술 개발과정과 PIM 구현방식을 통해 AI 연산의 효율성과 에너지 소비 개선방안을 탐구하는 기술 세션
HMSDK: 2-Tier Memory System Optimization Solution for DRAM and CXL Memory
Discover HMSDK의 2-Tier 메모리 시스템 최적화 솔루션과 CXL 메모리 활용을 위한 API 및 라이브러리 구현 방법을 통해 데이터센터 성능 향상 방안을 탐구합니다.
Emerging Memory Architecture Challenges with CXL Introduction
Explore CXL memory's impact on server systems through performance analysis of Meta's CacheLib and Hazelcast, examining bandwidth expansion and optimization strategies for next-gen memory architectures.
Managing Memory Tiers with CXL in Virtualized Environments
Explores hardware-managed memory tiering with CXL, introducing Intel® Flat Memory Mode and Memstrata allocator to optimize performance and reduce contention in virtualized cloud environments.
Introduction to CXL Technology and Memory Management Use Cases
Dive into CXL technology's evolution and practical applications, from memory expansion to distributed database caching and AI workload optimization, with insights from MemVerge's CEO.
Persistent CXL Memory Module with DRAM Performance - A Technical Overview
Explore CXL memory module innovations, focusing on DRAM-level performance, data persistence, and security enhancements for improved system architecture and power management solutions.
Foundation of GPU Programming
Explore GPU architecture, performance sources, and programming models. Gain insights into SIMD processing, DRAM parallelism, and software/memory hierarchies for confident GPU utilization.
SSD Architecture Challenges: Understanding DRAM Impact and Indirection Units
Dive into SSD architecture fundamentals, exploring DRAM impacts, Indirection Unit sizing, and architectural solutions for extreme-capacity storage systems with practical implementation strategies.
Data Redundancy and Scrubbing with RAID Offload Technology
Discover how RAID offload technology optimizes data redundancy solutions, reducing CPU and memory overhead while scaling performance with NVMe SSDs for efficient storage systems.
ZenHammer: Rowhammer Attacks on AMD Zen-based Platforms
Discover the first successful Rowhammer attack implementation on AMD Zen processors, exploring reverse engineering techniques, synchronization methods, and exploitation of DDR4/DDR5 memory vulnerabilities.
Go Go Gadget Hammer: Flipping Nested Pointers for Arbitrary Data Leakage
Explore a novel Rowhammer exploit technique targeting nested pointer dereferences in code patterns, demonstrating enhanced data leakage capabilities and improved attack primitives in Linux kernel security.
Yes, One-Bit-Flip Matters! Universal DNN Model Inference Depletion with Runtime Code Fault Injection
Explore FrameFlip, a groundbreaking attack method that depletes DNN model inference through single-bit fault injections, demonstrating universal effectiveness across major machine learning frameworks.