The NVIDIA A40 GPU is a breakthrough in data center performance, seamlessly combining professional graphics with powerful computing and AI acceleration. It addresses challenges in design, creativity, and science, leading the evolution of virtual workstations and server workloads. With cutting-edge features for tasks like ray-traced rendering and simulation, it empowers professionals with state-of-the-art capabilities for work anytime, anywhere.

The Power of NVIDIA Ampere Architecture

Products based on the NVIDIA Ampere architecture deliver notable improvements in performance, features, and benefits compared to previous-generation offerings, including:

NVIDIA Ampere Architecture CUDA Cores: Enhanced power efficiency and double-speed processing in single-precision floating-point (FP32) operations lead to substantial performance advancements in graphics and simulation workflows. This is particularly beneficial for intricate tasks like 3D computer-aided design (CAD) and computer-aided engineering (CAE).
Second-Gen RT Cores: Featuring up to twice the throughput compared to the prior generation, second-generation RT cores enable simultaneous execution of ray tracing alongside shading or denoising. This results in significant performance boosts for tasks such as photorealistic rendering in film or television production, architectural design assessments, and virtual prototyping of product designs. Additionally, this technology accelerates the rendering of ray-traced motion blur, delivering faster results with enhanced visual precision.
Third-Gen Tensor Cores: The introduction of Tensor Float 32 (TF32) precision brings about a remarkable up to 5x increase in training throughput compared to the prior generation, expediting AI and data science model training without necessitating any code modifications. With hardware support for structural sparsity, there is an extraordinary up to 10x throughput enhancement for inferencing. Additionally, Tensor Cores play a pivotal role in accelerating AI-enhanced graphics capabilities, including DLSS (Deep Learning Super Sampling), AI denoising, and advanced editing features in selected applications.
48GB GPU Memory: Equipped with ultra-fast GDDR6 memory, expandable to 96GB through NVLink, the A40 offers data scientists, engineers, and creative professionals the extensive memory needed for tasks involving massive datasets, such as ultra-high-resolution rendering, data science, and simulations.
Third-Gen NVIDIA NVLink: Link two A40 GPUs to scale from 48GB to 96GB of GPU memory. The increased GPU-to-GPU interconnect bandwidth of up to 112.5GB/s establishes a unified and scalable memory, enhancing graphics and compute workloads while handling larger datasets. The new compact NVLink connector allows deployment in a broader range of servers.
Virtualization-Ready: Enhanced with next-gen NVIDIA vGPU software, including the NVIDIA Virtual Workstation, the A40 supports larger and more powerful virtual workstation instances for remote users, facilitating high-end remote design, AI, and compute workloads.
PCI Express Gen 4: With PCIe Gen 4, the A40 doubles the bandwidth of PCIe Gen 3, boosting data-transfer speeds for CPU memory-intensive tasks like AI, data science, and 3D design. This faster PCIe performance also accelerates GPU Direct Memory Access (DMA) transfers, ensuring swift I/O communication of video data between the GPU and GPUDirect for Video-enabled devices - a robust solution for live broadcast. The A40 remains backward compatible with PCIe Express Gen 3 for deployment flexibility.
Data Center Efficiency and Security: Featuring a dual-slot, power-efficient design, the NVIDIA A40 is up to 2x as power-efficient as the previous generation, validated with a wide range of NVIDIA-certified systems from PNY-backed systems builders and integrators. Additionally, the NVIDIA A40 provides secure and measured boot with hardware root of trust capability, ensuring firmware integrity by preventing tampering or corruption.

NVIDIA A40 Specifications

GPU Memory	48GB GDDR6 with ECC (Error Correcting Code)
System Interface	PCIe Gen 4 x16 \| 31.5 GB/s Bidirectional
NVLink	2-way Low Profile (2-Slot) \| 112.5 GB/s Bidirectional
Display Outputs	3x DisplayPort 1.4
Max Power Consumption	300W
Form Factor	4.4” H x 10.5” L \| Dual Slot
Thermal Management	Passive
vGPU Software Support	NVIDIA vPC/vApps, NVIDIA RTX Virtual Workstation, NVIDIA Virtual Compute Server, NVIDIA AI Enterprise
vGPU Profiles Supported	Refer to the Virtual GPU Licensing Guide
NVEC \| NVDEC	1x \| 2x (includes AV1 decode)
Secure and Measured Boot	Yes \| With Hardware Root of Trust
NEBS Ready	Level 3
Power Connector	8-pin CPU

Professional Features

NVIDIA A40 integrates the performance and functionalities essential for extensive display experiences, virtual reality (VR), broadcast-grade streaming, and beyond.

Multi-Display Capabilities: Drive expansive CAVE Automatic Virtual Environments (CAVEs), video walls, virtual sets, broadcast, and location-based entertainment deployments with support for multiple 8K monitors. Utilize NVIDIA Mosaic multi-display technology featuring bezel correction and leverage NVIDIA's Warp and Blend SDK.
Quadro Sync Functionality: Synchronize multiple NVIDIA A40 GPUs with displays or projectors using NVIDIA Quadro Sync II. Create large-scale visualization environments, achieve artifact-free images across multiple displays with Frame Lock GPU outputs, or sync GPU outputs to an external timing source.
Video Encoding and Decoding: Benefit from dedicated video encoder (NVENC) and decoder engine (NVDEC) capabilities. Handle multiple streams simultaneously, expedite video exports, and support multi-stream applications for broadcast, security, and video serving.
Immersive VR Experiences: Powerfully drive Augmented Reality (AR) and Virtual Reality (VR) experiences on high-resolution Head-Mounted Displays (HMDs) with accelerated graphics and increased display bandwidth. Enable peak performance with four-way VR SLI, assigning 2 NVLink-connected A40 GPUs to each eye.
Enterprise-Grade Drivers and ISV Certifications: Virtual workstations, powered by NVIDIA RTX Virtual Workstation (vWS) software, leverage the same NVIDIA RTX platform as physical workstations. Benefit from extensive testing across diverse industry applications and over 100 Independent Software Vendor (ISV) certifications, ensuring optimal performance and stability.

Conclusion

The NVIDIA A40 stands as a testament to innovation, seamlessly combining powerful hardware with advanced features to address the challenges posed by modern workloads.

Now available through GreenNode, this high-end GPU is ready to empower your data centers with unparalleled performance, enabling you to tackle the most demanding tasks with confidence. Talk to our team if you have a specific use case!