#Hpc
5 posts in tag “Hpc”.
The Hidden Backbone of Parallelism: How Prefix Sums Power Distributed Computation
2025-09-21Discover how the humble prefix sum (scan) quietly powers GPUs, distributed clusters, and big data frameworks—an obscure but essential building block of parallel and distributed computation.
GPUDirect Storage in 2025: Optimizing the End-to-End Data Path
2025-09-16How modern systems move data from NVMe and object storage into GPU kernels with minimal CPU overhead and maximal throughput.
MPI vs. OpenMP in 2025: Where Each Wins
2025-07-04A practical guide to choosing between message passing and shared-memory parallelism for modern HPC and hybrid nodes.
Network Topologies for HPC: Fat-Trees, Dragonfly, Torus, and the Cost-Diameter-Bandwidth Optimization
2024-08-31A rigorous survey of HPC network topologies—fat-tree (InfiniBand), Dragonfly (Cray Cascade), torus (Blue Gene), Slim Fly—analyzing the fundamental tradeoffs in cost, diameter, bisection bandwidth, and fault tolerance.
Amdahl’s Law vs. Gustafson’s Law: What They Really Predict
2024-06-15When does parallelism pay off? Compare Amdahl’s and Gustafson’s models, see where each applies, and learn how to reason about speedups in practice.