Addressing Challenges of Cross-Architecture Dynamic Migration Over Heterogeneous Acceleration Systems
解决AI应用在不同加速硬件间平滑迁移的挑战。介绍统一抽象、跨架构编译和动态链接运行时等技术,实现无需改变开发习惯的跨系统应用迁移。
Unlocking Heterogeneous AI Infrastructure in Kubernetes Clusters - Leveraging HAMi
Explore HAMi for managing diverse AI devices in Kubernetes clusters. Learn about unified scheduling, GPU sharing, QoS, flexible strategies, and real-world applications for optimized resource utilization.
Metis - Fast Automatic Distributed Training on Heterogeneous GPUs
Efficient distributed training system for heterogeneous GPUs. Optimizes parallelism plans, improves resource utilization, and achieves significant speed-ups for large deep learning models.
Journey of Heterogeneous Computing in OpenInfra - From OpenStack Cyborg to AI Integration
Explore the 7-year evolution of OpenStack Cyborg's heterogeneous framework, from its community-driven inception to its role in AI infrastructure, and discover its impact on modern computing landscapes.
Meeting the Demands of Modern Datacenters - AI Computing and Efficiency with AMD
Discover how AMD's innovative solutions address modern datacenter challenges through AI integration, compute density, and power efficiency while maximizing performance and automation capabilities.
Memory Semantics and Data Movement with CXL and SDXI - Standards and Applications
Explore memory semantics and data movement standards through CXL and SDXI technologies, focusing on accelerator-assisted operations and heterogeneous computing architectures.
C++ for Multi-Accelerator, Multi-Vendor Systems - Meeting C++ 2022
Explore SYCL for C++ development on multi-accelerator systems. Learn basics, see demos, and discover how companies and institutions leverage this open-source, cross-device solution.
A Smooth Introduction to SYCL for C++20 Aficionados
Explore SYCL for cross-platform accelerator programming in C++20. Learn integration techniques, key features, and real-world applications in high-energy physics.
Why Do We Need a Compiler Architecture for Heterogeneous Computing
Explore compiler architecture for heterogeneous computing, addressing complex tasks and large data volumes. Learn about seamless integration between diverse computing units and platforms.
Legion: Programming Heterogeneous, Distributed Parallel Machines
Explore hierarchical data organization and computation in parallel programming, focusing on Legion's design for complex machine architectures.
Automating Distributed Heterogeneous Computing for Domain Experts
Explore new approaches for enabling domain experts to program future HPC systems efficiently, focusing on compiler and runtime techniques for distributed heterogeneous computing.
Heterogeneous Serverless Computing for High-Performance Computing and AI Workloads
Explore heterogeneous serverless computing for efficient HPC, HPDA, and AI workloads. Learn about innovative accelerator uses and flexible deployment models from edge to exascale.
Unlocking Heterogeneous AI Infrastructure K8s Cluster - Leveraging the Power of HAMi
探索HAMi项目如何优化异构AI设备在K8s集群中的管理,提高资源利用率,支持灵活调度策略,并解决实际生产环境中的挑战。
Parallelism Safety-Critical Guidelines for C++
Explore parallelism safety guidelines for C++, covering challenges in multicore architectures and potential issues with newer features. Gain insights into future rules for safety-critical software development.
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.
Heterogeneous Multi-Core Systems for Efficient Edge Machine Learning
Explore heterogeneous multi-core systems for efficient embedded ML, addressing challenges in core optimization, workload mapping, and data sharing. Gain insights from KULeuven MICAS research.