NVIDIA Ada Lovelace Architecture

Experience fast, interactive performance—powered by the latest NVIDIA Ada Lovelace architecture-based GPU—with ultra-fast, onboard graphics memory technology and optimized software drivers for professional applications. The Shader Execution Reordering (SER) system allows on-the-fly organization and reordering of workloads, grouping similar performing threads so the streaming multiprocessor (SM) and RT Core can operate more efficiently.

CUDA Cores

The NVIDIA Ada Lovelace architecture-based CUDA® cores deliver up to 2X the single-precision floating-point (FP32) throughput compared to the previous generation, providing significant performance improvements for graphics workflows such as 3D model development and compute for workloads like desktop simulation for CAE.

Third-Generation RT Cores

Third-generation RT Cores provide up to 2X the throughput of the previous generation and the ability to run ray tracing with either shading or denoising capabilities concurrently. This accelerates renders for M&E content creation, AECO design evaluations, and manufacturing virtual prototyping. Third-generation RT Cores deliver up to 2X the ray-tracing performance over the previous generation, delivering groundbreaking performance for photorealistic rendering. Enhanced RT Cores combined with new SER technology dynamically reorder inefficient workloads, dramatically improving shader performance to accelerate end-to-end ray-traced image rendering performance.

Fourth-Generation Tensor Cores

Fourth-generation Tensor Cores provide up to 5X faster AI training performance than the previous generation with FP16 precision. Support for the new FP8 data format for inference provides more than 5X faster performance than the previous generation and reduces data memory usage by half (compared to the FP16 data format).

Encode and Decode Engines

RTX 5000 includes two video encode engines and two decode engines, including support for the AV1 video format and the performance required for multi-stream video applications for security and video serving.

PCIe Gen 4

 RTX 5000 supports PCIe Gen4, which doubles the bandwidth of PCIe Gen3 from 15.75GB/s to 31.5GB/s for x16 connections, improving data transfer speeds from CPU memory for data-intensive tasks such as AI, data science, and creating large 3D models and scenes. Faster PCIe performance also accelerates GPU DMA transfers, providing faster video data transfers from GPUDirect for Video-enabled devices and faster IO with GPUDirect Storage.