The NVIDIA RTX and Data Center GPU Benchmarks for Deep Learning whitepaper reviewed by PNY and NVIDIA, but developed and published by EXXACT, takes a careful and nuanced look at ResNet-50, a popular means of measuring the performance of machine learning (ML/AI) accelerators. NVIDIA's Data Center GPUs were tested with the Amber 22 GPU benchmark. Since Amber is a popular molecular dynamics application it gives real-world context to the outstanding benefits GPU acceleration offers for the scientific and technical computing community.
"Is the GPU the New CPU?" is the latest e-guide written by Kurt Cagle from Data Science Central which looks at how GPUs are starting to overshadow their older CPUs siblings as the powerhouse of the modern computer, handling complex neural networks and visualization tasks. Other topics covered include "Are DPUs the Next Iteration of GPUs? and "GPU Networks and the Future". Download the E-Guide below.
3D workflows are now an essential component of every engineering and manufacturing discipline. The rise of remote work, globally spread teams, an explosion in the number of often incompatible software tools, and the demand for compute-heavy technologies makes design collaboration exponentially harder – especially now. NVIDIA Omniverse™ Enterprise depends on NVIDIA RTX technology (RTX A6000, RTX A5500, RTX A5000, RTX A4500 and RTX A4000) – to let geographically disparate groups or hybrid office and work from home teams transparently exchange 3D data between popular 3D design tools they already know, maintain a single source of truth through sophisticated publish and subscribe capabilities, protect sensitive IP, and realize real-time multi-participant design collaboration and reviews, simulation, even digital twins of entire physical facilities.