NVIDIA DRIVE AI Solutions
…platform delivering industry-leading performance. By combining raw computing power with trusted DRIVE AGX Orin and DRIVE AGX Thor ecosystem partners, we provide a seamless transition to production. DRIVE AGX is powered…
Search
Top stories
Automating Inference Optimizations with NVIDIA TensorRT LLM AutoDeploy | NVIDIA Technical Blog
…For the benchmarking scenario in Figure 1 (ISL/OSL=8k/16k), Nemotron-Flash outperforms Qwen2.5 highlighting how novel model architectures can be quickly onboarded to achieve production-ready performance. Data was…
Accelerated X-Ray Analysis for Nanoscale Imaging (XANI) of Novel Materials | NVIDIA Technical Blog
…Through extensive experimentation on the latest GPUs and high-performance Lustre storage systems, three critical optimizations were performed to achieve peak I/O performance: GDS, multithreaded HDF5, and data layout (details to…
Introducing NVIDIA BlueField-4-Powered CMX Context Memory Storage Platform for the Next Frontier of AI | NVIDIA Technical Blog
…Einav Zilberstein Einav Zilberstein is a senior product manager for NVIDIA DOCA storage software, helping data centers and AI customers adopt DPU‑accelerated, high‑performance, and secure storage networking solutions. Einav has…
Using NVFP4 Low-Precision Model Training for Higher Throughput Without Losing Accuracy | NVIDIA Technical Blog
…NVIDIA NeMo Megatron Bridge provides production-ready low-precision training recipes that allow seamless switching between precision formats, supporting efficient large-scale model training with minimal code modifications. AI-generated content may…
NVIDIA Vera Rubin POD: Seven Chips, Five Rack-Scale Systems, One AI Supercomputer | NVIDIA Technical Blog
…POD. Third-generation NVIDIA MGX rack-scale architecture Production-grade AI racks must excel across several critical areas: rapid time to volume, proven performance at scale, deep hardware-software co-design, resiliency…
NVIDIA JetPack Software Stack
…With full support for NVIDIA Jetson platforms, JetPack 7 provides ultra-low latency, deterministic performance, and scalable deployment for machines that interact with the physical world. JetPack 7 Overview JetPack 7 gives…
Integrate Physical AI Capabilities into Existing Apps with NVIDIA Omniverse Libraries | NVIDIA Technical Blog
…Step the sensor simulation and retrieve rendered outputs products = renderer.step(render_products={"/Render/Product_Robot_01"}, delta_time=1.0/60) # 4. Save rendered output into PNG via numpy DLPack and…
Making Softmax More Efficient with NVIDIA Blackwell Ultra | NVIDIA Technical Blog
…nvcc -O3 -gencode=arch=compute_103a,code=sm_103a --extended-lambda -o /tmp/exp2-gb300.out exp2-gb300.cu Sample results We see that GB300 performs about 2x higher in FLOPs performance…
Run High-Throughput Reinforcement Learning Training with End-to-End FP8 Precision | NVIDIA Technical Blog
…To make these workloads viable, researchers and engineers are turning to low-precision datatypes like FP8 to boost performance in training and throughput-oriented generation. Moreover, in some scenarios where generation is…