The auto-updated technology brief
NVIDIA introduces Nemotron 3 Ultra using NVFP4 4-bit floating point quantization with Model Optimizer to efficiently move large model weights for longer context windows.
NVIDIA Omniverse NuRec is a neural reconstruction pipeline that builds high-fidelity 3D representations of real-world environments from multisensor data, optimized using NVIDIA Nsight developer tools.
NVIDIA's GPU-accelerated query engines overcome memory and I/O bandwidth constraints using hardware advances like high bandwidth memory, NVLink-C2C, and decompression engines in the GB200 NVL4.
NVIDIA introduces Isaac GR00T to streamline humanoid robot policy development, addressing fragmented workflows and infrastructure challenges.
Training large language models (LLMs) at massive scale faces infrastructure challenges due to long runtimes and thousands of GPUs. Nonuniform tensor parallelism helps mitigate slowdowns caused by device unavailability and resource fluctuations.
The Open Secure AI Alliance brings together industry leaders to enhance AI safety and security by leveraging open source software, which supports critical sectors like cloud computing and cybersecurity.
Amazon EC2 C9g and C9gd instances, powered by AWS Graviton5, are now generally available. They deliver up to 25% better compute performance than previous generation, feature fastest memory in the cloud, and offer local NVMe storage options for demanding workloads.
AWS announces Claude Sonnet 5 availability, Amazon WorkSpaces for AI agents GA, OpenSearch optimization for log analytics, SageMaker AI scaling improvements, and key AWS service availability changes.
NVIDIA researchers have developed an Ising decoding method that reduces logical error rates in quantum color codes by over 300 times, advancing fault-tolerant quantum computing.
NVIDIA released Video Codec SDK 13.1 featuring zero-copy transcode, AV1 B-frames support, and frame-accurate seeking to enhance video pipeline efficiency and quality.
NVIDIA's Vera CPU enhances AI factory throughput by accelerating agentic workloads that involve multi-step workflows combining inference, tool use, and orchestration, improving CPU performance between model steps.
NVIDIA's NeMo framework addresses data limitations in financial NLP by generating synthetic data to improve LLM fine-tuning for trading research, risk modeling, and surveillance.