NVIDIA Ampere Architecture

NVIDIA RTX A2000 is the most powerful dual slot, low profile GPU solution offering high performance real-time ray tracing, AI-accelerated compute, and professional graphics rendering, all of this in a compact design with unmatched performance. Building upon the major SM enhancements from the Turing GPU, the NVIDIA Ampere architecture enhances ray tracing operations, tensor matrix operations, and concurrent executions of FP32 and INT32 operations.

CUDA Cores

The NVIDIA Ampere architecture-based CUDA cores bring up to 2.7X 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 such as desktop simulation for computer-aided engineering (CAE). The RTX A2000 enables two FP32 primary data paths, doubling the peak FP32 operations.

Second Generation RT Cores

Incorporating 2nd generation ray tracing engines, NVIDIA Ampere architecture-based GPUs provide incredible ray traced rendering performance. For the first time NVDIA is introducing RT Cores into a low profile form factor GPU. A single RTX A2000 board can render complex professional models with physically accurate shadows, reflections, and refractions to empower users with instant insight. Working in concert with applications leveraging APIs such as NVIDIA OptiX, Microsoft DXR and Vulkan ray tracing, systems based on the RTX A2000 will power truly interactive design workflows to provide immediate feedback for unprecedented levels of productivity. The RTX A2000 is more than 5X faster in ray tracing compared to the previous generation. This technology also speeds up the rendering of ray-traced motion blur for faster results with greater visual accuracy.

Third Generation Tensor Cores

Purpose-built for deep learning matrix arithmetic at the heart of neural network training and inferencing functions, the RTX A2000 includes enhanced Tensor Cores that accelerate more datatypes and includes a new Fine-Grained Structured Sparsity feature that delivers more than 2X throughput for tensor matrix operations compared to the previous generation. New Tensor Cores will accelerate two new TF32 and BFloat16 precision modes. Independent floating-point and integer data paths allow more efficient execution of workloads using a mix of computation and addressing calculations.

PCIe Gen 4

The RTX A2000 supports PCI Express Gen 4, which provides double the bandwidth of PCIe Gen 3, improving data-transfer speeds from CPU memory for data-intensive tasks like AI and data science.

Higher Speed GDDR6 Memory

Built with 6GB GDDR6 memory delivering up to 2X greater throughput for ray tracing, rendering, and AI workloads than the previous generation. The RTX A2000 provides ideal memory footprint to address the datasets and models in latency-sensitive professional applications and at volume.

Error Correcting Code (ECC) on Graphics Memory

Meet strict data integrity requirements for mission critical applications with uncompromised computing accuracy and reliability for workstations.

Fifth Generation NVDEC Engine

NVDEC is well suited for transcoding and video playback applications for real-time decoding. The following video codecs are supported for hardware-accelerated decoding: MPEG-2, VC-1, H.264 (AVCHD), H.265 (HEVC), VP8, VP9, and AV1.

Seventh Generation NVENC Engine

NVENC can take on the most demanding 4K or 8K video encoding tasks to free up the graphics engine and the CPU for other operations. The RTX A2000 provides better encoding quality than software-based x264 encoders.

Graphics Preemption

Pixel-level preemption provides more granular control to better support time-sensitive tasks such as VR motion tracking.

Compute Preemption

Preemption at the instruction-level provides finer grain control over compute tasks to prevent long- running applications from either monopolizing system resources or timing out.

NVIDIA RTX IO

Accelerating GPU-based lossless decompression performance by up to 100x and 20x lower CPU utilization compared to traditional storage APIs using Microsoft’s new DirectStorage for Windows API. RTX IO moves data from the storage to the GPU in a more efficient, compressed form, and improving I/O performance.

NVIDIA Mosaic Technology

Transparently scale the desktop and applications across up to 4 GPUs and 16 displays from a single workstation while delivering full performance and image quality.

DisplayPort 1.4a

Support up to four 5K monitors at 60Hz, or dual 8K displays at 60Hz per card. The RTX A2000 supports HDR color for 4K at 60Hz for 10/12b HEVC decode and up to 4K at 60Hz for 10b HEVC encode. Each DisplayPort connector can drive ultra-high resolutions of 4096x2160 at 120 Hz with 30-bit color.

NVIDIA RTX Desktop Manager

Gain unprecedented end-user control of the desktop experience for increased productivity in single large display or multi-display environments, especially in the current age of large, widescreen displays.

OpenGL Quad Buffered Stereo Support

Provides a smooth and immersive 3D Stereo experience for professional applications.

Ultra-High-Resolution Desktop Support

Get more Mosaic topology choices with high resolution displays devices with a 32K Max desktop size.

Professional 3D Stereo Synchronization

Robust control of stereo effects through a dedicated connection to directly synchronize 3D stereo hardware to an NVIDIA RTX professional graphics card.

NVIDIA RTX Experience

NVIDIA RTX Experience delivers a suite of productivity tools to your desktop workstation, including desktop recording in up to 8K, automatic alerts for the latest NVIDIA RTX Enterprise driver updates, and access gaming features. The application is available for download.

Software Optimized for AI

Deep learning frameworks such as Caffe2, MXNet, CNTK, TensorFlow, and others deliver dramatically faster training times and higher multi-node training performance. GPU accelerated libraries such as cuDNN, cuBLAS, and TensorRT delivers higher performance for both deep learning inference and High- Performance Computing (HPC) application.

NVIDIA CUDA Parallel Computing Platform

Natively execute standard programming languages like C/C++ and Fortran, and APIs such as OpenCL, OpenACC and Direct Compute to accelerates techniques such as ray tracing, video and image processing, and computation fluid dynamics.

Unified Memory

A single, seamless 49-bit virtual address space allows for the transparent migration of data between the full allocation of CPU and GPU memory.

NVIDIA Enterprise-Management Tools

Maximize system uptime, seamlessly manage wide-scale deployments, and remotely control graphics and display settings for efficient operations.

Specification 

Architecture NVIDIA Ampere Architecture

Process Size 8nm

Transistors 13.25 Billion

Die Size 276 mm2

CUDA Cores 3328

Tensor Cores 104

RT Cores 26

Peak FP32 Performance 7.99 TFLOPS

RT Core Performance 15.6 TFLOPS

Tensor Core Performance 63.9 TFLOPS

GPU Memory 6 GB GDDR6 ECC

GPU Memory Clock 6001 MHz

Memory Interface 192-bit

Memory Bandwidth 288 GB/sec

System Interface PCI Express 4.0 x16

Display Connectors 4x mDP Latching

Form Factor 2.713" H x 6.6" Low Profile | Dual Slot

Thermal Solution Active Fansink

Max Digital Resolution 2x 7680 x 4320 at 60 Hz

Max Displays 4x at 5K | 5120 x 2880 at 60 Hz

HDR Support Yes

AR | VR | XR Ready Yes

Quad Buffered Stereo Yes

Hardware Accelerated BVH Yes

NVIDIA DLSS Yes

NVIDIA Mosaic, Warp and Blend Yes

NVIDIA RTX Experience Yes

NVIDIA RTX Desktop Manager Yes

NVIDIA Broadcast Application Yes

Graphics APIs DirectX 12.07, Shader Model 5.17, Open GL 4.68, Vulkan 1.2

Compute APIs CUDA, DirectCompute, Open CL

NVENC | NVDEC - 1x H.264 and HVEC Encoders - 1x Decode Engine