Learning Local Equivariant Representations for Large-Scale Atomistic Dynamics
Explore Allegro, a local equivariant deep learning interatomic potential for efficient, accurate, and scalable atomistic dynamics simulations, including applications in biomolecular systems.
De/Re-Composition of Data-Parallel Computations via Multi-Dimensional Homomorphism - PLDI 2024
Explore a systematic approach for efficient (de/re)-composition of data-parallel computations using Multi-Dimensional Homomorphisms, applicable to various architectures and outperforming state-of-the-art solutions.
Heterogeneous Concurrency - A New Frontier for Weak Memory
Explore heterogeneous processors' shared memory interfaces and their impact on weak memory concurrency models in emerging computing domains.
Evolving Weak Memory Models for Evolving Architectures
Explore evolving weak memory models for modern architectures, focusing on advancements and challenges in concurrent programming and hardware design.
System-Level Weak Memory Models: The Need for Formalisation, ISA Semantics Integration and Model Diversity
Explore system-level weak memory models, focusing on formalization, ISA semantics integration, and model diversity in computer architecture.
Programmers Love Mind-Bogglingly Complicated Weak Memory Models
Explore mind-boggling complexities of weak memory models in programming, delving into their intricacies and implications for software development.
Designing Wait-free Weak Reference Counting
Explore wait-free weak reference counting design techniques for efficient memory management in concurrent systems.
Vectorized Query Execution with SIMD - Lecture 8
Explore advanced vectorized query execution techniques using SIMD for optimized database performance in this comprehensive lecture by Prof. Andy Pavlo.
Scalable Multi-Node AI Workloads in Multi-Tenant AI Clouds Using SDN K8s Networking
Optimize AI workload performance in multi-tenant clouds using SDN K8s networking. Learn techniques for network isolation, efficient bandwidth, and QoS with Open Source solutions like OVS, OVN, and OVN-Kubernetes CNI.
Deploying LLM Workloads on Kubernetes Using WasmEdge and Kuasar
Deploy LLM workloads efficiently on Kubernetes using WasmEdge and Kuasar. Explore solutions for faster container startup, reduced management overhead, and enhanced scalability in LLM operations.
Principles and Practice of Scalable and Distributed Deep Neural Networks Training and Inference
Dive into scalable and distributed deep neural networks, exploring advanced training and inference techniques for enhanced performance and efficiency in modern AI systems.
Networking Technologies for High-Performance Computing - Principles and Solutions
Explore advanced networking principles and solutions for high-performance computing systems, focusing on interconnect technologies, design strategies, and optimization techniques.
High Performance Machine Learning, Deep Learning, and Data Science - Principles and Practice
Master high-performance computing principles for machine learning and deep learning, focusing on practical implementation strategies and advanced techniques for optimizing data science workflows.
Ray - A Distributed System for Emerging AI Applications
Explore Ray, a scalable distributed system designed for AI workloads, featuring task-parallel and actor abstractions, in-memory storage, and a distributed scheduler for emerging AI applications.
Servo - Designing and Implementing a Parallel Browser
Explore Servo, an experimental browser engine designed for multi-core hardware, focusing on parallelization and security through Rust. Learn about its goals, design, and potential impact on web browsing.
How Do I Compute the Body Electric?
Explore computational physics and supercomputing to revolutionize personalized medicine through advanced simulations of the human body.
Near Optimal Massively Parallel Graph Connectivity
Explore efficient parallel algorithms for graph connectivity, focusing on near-optimal solutions and innovative techniques for large-scale graph processing.
Parallel Reachability in Almost Linear Work and Square Root Depth
Explore efficient parallel algorithms for graph reachability, focusing on distributed computing and improved runtime complexity in large-scale networks.
Game Server Performance on Tom Clancy's The Division 2
Explore technical challenges and solutions for game server performance in Tom Clancy's The Division 2, focusing on parallelization and multithreading techniques for improved efficiency.
GPU Security Exposed
Explore GPU security vulnerabilities, including shared memory risks and privilege escalation issues, with insights on protecting popular devices from potential exploits.
Intrinsics - Low-Level Engine Development with Burst - Unite Copenhagen
Explore low-level engine development using Burst Compiler's expanded capabilities, including CPU intrinsics, SIMD optimization, and performance comparisons with C++.