Enabling high-performance artificial intelligence with NVIDIA GPUs
GPUs are important technology building blocks for accelerated computing and artificial intelligence due to their inherent parallel processing technology, allowing them to execute multiple tasks or processes simultaneously. NVIDIA has taken this further with its latest generation of GPUs, which feature Tensor Cores. These cores, like standard graphics processing cores, allow mixed precision computations, making them an efficient use of the available processing power for AI and signal processing.
GPU Processing Solutions from Curtiss-Wright
Curtiss-Wright works closely with its technology partners, NVIDIA and WOLF Advanced Technology, to offer rugged, high-performance SOSA® aligned NVIDIA GPU-based cards and subsystems. Customers pair NVIDIA’s GPU technologies with Curtiss-Wright’s portfolio of CPU and FPGA products to complete their system-level solutions.
NVIDIA's latest generation technologies provide enhanced performance while reducing power consumption and leverage NVIDIA's extensive collection of AI tools and workflows. Curtiss-Wright's NVIDIA-based solutions include:
- processing capabilities with rugged mission computing
- processing languages/frameworks such as NVIDIA CUDA™ and OpenCL
- modules for video output, image, and data processing, video encoding to HEVC (H.265) and AVC (H.264), and high-performance embedded compute (HPEC) tasks
- GPU-based modules also include Ada, Hooper, and Blackwell GPUs
- SOSA® aligned 3U and 6U form factors
GPU Processing Solutions from Curtiss-Wright
Curtiss-Wright partners with NVIDIA and WOLF Advanced Technology to select GPU processors for demanding long lifecycle embedded computing applications. Curtiss-Wright selects industrial qualified processors to ensure high-reliability in rugged VPX, VME, and XMC form factor cards and systems based on open standards like MOSA, SOSA, and CMOSS. Curtiss-Wright solutions address customer's sensor processing, artificial intelligence, and machine learning requirements.
