NumPy Programming Tutorials & Engineering Articles

NVIDIA nvmath-python is a library designed to bridge the gap between the Python scientific community and NVIDIA CUDA-X math libraries. It gives Python users…

This second installment explores how Ray’s higher-level libraries—Serve, Data, and Train—abstract the complexities of running AI workloads on Google's TPU slices. Ray Serve uses a simple topology configuration to correctly gang-schedule large multi-host models, while Ray Data eliminates data-loading bottlenecks by feeding accelerators directly with native JAX batches. Finally, JaxTrainer streamlines distributed training across TPUs by automatically handling cross-slice coordination, checkpointing, and fault tolerance.

Developers building 3D, design, simulation, robotics, and industrial digital twin applications need ways to bring physical AI capabilities into the tools and…

Distributed AI training is notoriously fragile because losing a single machine typically crashes the entire multi-node job, forcing a time-consuming, full-workload infrastructure restart. To address this, Google’s JAX ecosystem utilizes elastic training via Pathways, which converts a hardware failure into a catchable Python exception so the running process can survive. When an unplanned failure occurs, the system automatically replaces only the broken worker, restores the last viable checkpoint from Cloud Storage, and resumes training in place—minimizing total downtime to under two minutes without ever restarting the main controller process.

A massive-scale X-ray free-electron laser (XFEL) enables tracking structural and electron dynamics in novel systems, including fusion materials, semiconductors…

In a previous post, we introduced the Universal Sparse Tensor (UST), enabling developers to decouple a tensor’s sparsity from its memory layout for greater…

Physical AI—AI systems that perceive, reason, and act in physically grounded simulated environments—is changing how teams design and validate robots and…

Computer-aided engineering (CAE) is shifting from human-driven workflows toward AI-driven ones, including physics foundation models that generalize across…

The article discusses how accelerated computing, particularly through NVIDIA's technologies, is transforming scientific experiments at large research facilities like the NSF-DOE Vera C.

The article discusses the development of chDB, a Python library that integrates ClickHouse with Pandas DataFrames for high-performance SQL querying.

The 2025 Typed Python Survey provides insights into the adoption of Python's type system, highlighting code quality and flexibility as primary motivations for its use.

The article discusses how to simulate an accurate radio environment for 5G and 6G systems using the NVIDIA Aerial Omniverse Digital Twin (AODT).

The article discusses the launch of NVIDIA CUDA Tile with CUDA 13. 1, which introduces a virtual instruction set for tile-based parallel programming.

The article discusses enhancing robot perception efficiency on the NVIDIA Jetson Thor platform by utilizing specialized hardware accelerators alongside powerful GPUs.

This article discusses the integration of the Newton physics engine with NVIDIA Isaac Lab for training quadruped locomotion policies and simulating cloth manipulation.

The article discusses Autodesk Research's development of the Accelerated Lattice Boltzmann (XLB) library, which enhances computational fluid dynamics (CFD) performance using NVIDIA's Warp and GH200...

The article discusses the integration of AI-powered simulations in computer-aided engineering (CAE) to accelerate design processes.

The article discusses the significant updates in CUDA Toolkit 13.

The article discusses the increasing adoption of JAX in robotics, highlighting its efficiency in optimal control and simulation. It features insights from Max Muchen Sun, a Robotics Ph. D.

The article discusses the advancements in single-cell analysis facilitated by RAPIDS-singlecell, an open-source tool that leverages GPU acceleration to handle large datasets efficiently.

The article discusses the enhancements in the Forest Inference Library (FIL) within NVIDIA cuML 25. 04, focusing on its capabilities for fast inference of tree-based models.

The article discusses how to accelerate Deep Learning (DL) and Large Language Model (LLM) inference using Apache Spark in cloud environments.

NVIDIA cuPyNumeric 25. 03 is a fully open-source library designed as a drop-in replacement for NumPy, leveraging the Legate framework for accelerated computing.

The article discusses advancements in Federated Learning (FL) specifically in the context of large language models (LLMs), focusing on the challenges of communication overhead and memory constraint...

NVIDIA cuML has introduced a zero code change capability that allows data scientists and machine learning engineers to accelerate scikit-learn applications on NVIDIA GPUs without modifying existing...

The article discusses how to leverage GPU acceleration for algorithmic trading simulations using Numba, highlighting the significant performance improvements achievable—over 100x faster simulations.

The article discusses how NAVER Place optimizes its small language model (SLM)-based vertical services using NVIDIA TensorRT-LLM, enhancing usability and performance.

This article explores the various input formats supported by ClickHouse for data ingestion, focusing on performance and efficiency.

The article discusses the importance of GPU acceleration in data science, highlighting how NVIDIA's RAPIDS suite can significantly enhance performance for data processing tasks.

The article introduces tile-based programming in Warp 1. 5. 0, highlighting new Python primitives that enhance GPU programming efficiency.

The article introduces NVIDIA cuPyNumeric, an accelerated and distributed implementation of the NumPy API that allows users to scale their NumPy programs seamlessly from laptops to supercomputers w...

The NVIDIA RAPIDS v24.

Federated learning is transforming the development of autonomous vehicles (AVs) by allowing decentralized training using locally collected data.

The article discusses the latest features and improvements in NVIDIA CUDA-Q v0. 8, an open-source programming model for hybrid quantum-classical applications.

The article discusses the importance of securing AI model files against unauthorized access and introduces the concept of canaries as a detection mechanism.

The article discusses how to build a zero-copy AI sensor processing pipeline using OpenCV within the NVIDIA Holoscan SDK.

The article discusses WholeGraph, a feature in the RAPIDS cuGraph library designed to optimize memory and retrieval for Graph Neural Networks (GNNs).

The article discusses how Meta has improved machine learning model training times through the implementation of Lazy Imports and the Python Cinder runtime.

The article discusses the challenges of drug discovery and introduces NVIDIA BioNeMo's new models, MolMIM and DiffDock, which enhance molecule generation and molecular docking.

The article discusses how NVIDIA Holoscan is being utilized to accelerate ptychography workflows at the Diamond Light Source, a leading synchrotron facility.

Spotify has introduced Voyager, a new nearest-neighbor search library that significantly improves upon its predecessor, Annoy, by offering increased speed and accuracy.

NVIDIA's Differentiable Slang is a new shading language designed to unify real-time, inverse, and differentiable rendering, enabling seamless integration of machine learning with graphics programmi...

NVIDIA hosted a two-day training session at Black Hat USA 2023, focusing on the unique security risks associated with machine learning (ML).

The article discusses the security vulnerabilities found in machine learning research code, particularly focusing on the analysis of the Meta Kaggle for Code dataset.

The article discusses the release of MediaPipe Solutions for Raspberry Pi and iOS, highlighting the capabilities of the new iOS SDK and the updated Python SDK for Raspberry Pi.

The article discusses the importance of time series data in observability at Pinterest, detailing the development of TScript, a domain-specific language designed to manipulate time series data effi...

The article discusses the advancements in single-cell RNA sequencing analysis using the RAPIDS-singlecell library, which leverages GPU acceleration to significantly enhance performance.

The article discusses the integration of distributed deep learning with Apache Spark 3. 4, highlighting new built-in APIs for both distributed model training and inference.

The article discusses RAPIDS RAFT, a library designed to optimize machine learning and data analytics on GPUs by providing reusable computational patterns.