The NVIDIA RTX™ 6000 Ada Generation represents the **most powerful workstation GPU** boasting exceptional real-time ray tracing, AI-accelerated processing, and top-tier graphics rendering capabilities. With advancements in Streaming Multiprocessors (SM) from the Ada Lovelace GPU, it features increased core counts, higher clock speeds, and an expanded L2 cache, enhancing ray tracing, tensor matrix computations, and frame rates through DLSS 3.0.

The CUDA Cores in the NVIDIA Ada Lovelace architecture deliver over **2X** the single-precision floating point (FP32) throughput of prior generations, significantly boosting graphics workflows for 3D modeling and desktop simulation in computer-aided engineering (CAE). The RTX 6000 enhances FP32 operations by enabling dual primary data paths.

Equipped with advanced ray tracing engines, the NVIDIA Ada Lovelace-based GPUs deliver remarkable ray-traced rendering performance. The RTX 6000 enables users to render intricate professional models with precision shadows, reflections, and refractions, integrating seamlessly with APIs like NVIDIA OptiX, Microsoft DXR, and Vulkan ray tracing for unmatched productivity and immediate design feedback, achieving up to **2X** faster ray-triangle intersection throughput compared to predecessors.

Designed for deep learning operations, the RTX 6000 features enhanced Tensor Cores, which further improve data type acceleration and maintain Fine-Grained Structured Sparsity, allowing for over **2X** throughput for tensor matrix tasks versus the previous generation. The upgraded Tensor Cores also support new FP8 precision modes, optimizing workflow efficiency.

Built with **48GB GDDR6 memory**, the RTX 6000 offers up to **25%** higher throughput for ray tracing, rendering, and AI workloads compared to its predecessor, providing an expansive graphics memory footprint for handling the largest and most latency-sensitive datasets in professional applications.

With support for **PCI Express Gen 4**, the RTX 6000 doubles the bandwidth of previous iterations, significantly enhancing data transfer speeds from CPU memory for demanding tasks such as AI and data-centric applications.

Achieve stringent data integrity for critical applications with unmatched computing precision and reliability tailored for high-performance workstations.

NVDEC is optimized for transcoding and video playback needs in real-time decoding applications, supporting key video codecs such as MPEG-2, VC-1, H.264 (AVCHD), H.265 (HEVC), VP8, VP9, and AV1 formats, allowing 8K/60 video encoding for professional editing tasks.

NVENC handles demanding **4K** and **8K video encoding** tasks efficiently, allowing the graphics engine and CPU to serve other operations. The RTX 6000 features superior quality encoding compared to software-based x264 encoders and incorporates **AV1 video encoding**, which is 40% more efficient than **H.264** for **4K HDR** video, achieving better quality at the same bitrate.

Pixel-level preemption offers enhanced control, providing support for time-sensitive processes like VR motion tracking.

Instruction-level preemption delivers improved control over compute tasks, preventing long-running applications from monopolizing system resources or timing out.

Support for up to **four 5K monitors** at **60Hz**, or **dual 8K displays** at **60Hz** per card. The RTX 6000 also supports HDR color for **4K @ 60Hz** for **10/12b HEVC** decode and up to **4K @ 60Hz** for HEVC encoding. Each DisplayPort connector can drive ultra-high resolutions of **4096x2160 @ 120 Hz** with **30-bit color**.

Achieve unparalleled user control over desktop functionality to enhance productivity in both single large displays and multi-display environments, optimized for contemporary wide-format screens.

Synchronize display and image output for up to **32 displays** from **8 GPUs**, connected through two Sync II boards within a single system, significantly reducing the hardware footprint required for advanced video visualization setups.

Each frame lock connector features a self-locking retention mechanism to maintain a robust connection with the frame lock cable, enhancing connectivity and maximizing productivity.

Deliver an immersive 3D Stereo experience, ideal for professional applications requiring smooth performance.

Expand Mosaic topology options with high-resolution display devices supporting a **32K Max** desktop size.

Achieve robust control over stereo effects through a dedicated connection that directly synchronizes 3D stereo hardware with an NVIDIA RTX professional graphics card.

NVIDIA's virtual GPU (vGPU) software empowers a standard workstation to be transformed into multiple high-performance virtual workstation instances, enabling **remote users** to efficiently share resources for demanding design, AI, and computational tasks.

The NVIDIA RTX Experience provides a comprehensive suite of productivity tools for your workstation, featuring desktop recording in up to **8K**, automatic notifications for the latest NVIDIA RTX Enterprise driver updates, and diverse gaming features. Download the application here.

Leading deep learning frameworks such as Caffe2, MXNet, CNTK, and TensorFlow realize dramatically accelerated training times and superior multi-node performance. GPU-accelerated libraries like cuDNN, cuBLAS, and TensorRT enhance both deep learning inference and High-Performance Computing (HPC) tasks.

Directly execute popular programming languages such as C/C++ and Fortran, alongside APIs like OpenCL, OpenACC, and Direct Compute, to accelerate processes such as ray tracing, video/image processing, and computational fluid dynamics.

Utilize a cohesive, **49-bit** virtual address space that facilitates seamless data migration between complete allocations of CPU and GPU memory.

GPUDirect technology accelerates communication between the GPU and video I/O devices by circumventing the need for superfluous system memory copies and minimizing CPU load.

Maximize system uptime while efficiently managing extensive deployments and remotely controlling graphics and display settings for optimal operations.