NVIDIA L40S Tensor Core GPU 48GB GDDR6 AI VDI Graphics 900-2G179-0000-000: Unleash Unprecedented Performance
Revolutionary GPU Architecture: NVIDIA Ada Lovelace
The NVIDIA L40S Tensor Core GPU 48GB GDDR6 AI VDI Graphics 900-2G179-0000-000 is powered by the cutting-edge NVIDIA Ada Lovelace architecture. This advanced architecture ensures unparalleled performance and efficiency, making it the ideal choice for IT professionals seeking top-tier solutions for AI, machine learning, and virtual desktop infrastructure (VDI) applications.Exceptional Memory and Speed
Equipped with 48GB of GDDR6 memory, the NVIDIA L40S delivers lightning-fast data processing capabilities. The base clock speed of 1065 MHz and a maximum boost clock speed of 2520 MHz ensure that your applications run smoothly and efficiently, even under the most demanding workloads.Unmatched Computational Power
With 568 Tensor Cores and 18176 CUDA Cores, the NVIDIA L40S offers extraordinary computational power. It boasts a single-precision performance of 91.6 TF, mixed-precision performance of up to 366.4 TF*, INT8 performance of up to 1466 TOPS*, and INT4 performance of up to 1466 TOPS*. This makes it an exceptional choice for deep learning, AI, and other compute-intensive tasks.Advanced Connectivity and Compatibility
The NVIDIA L40S features a PCIe 4.0 x16 interface, ensuring high-speed data transfer and compatibility with a wide range of server platforms. With 4 Display Ports, it supports multiple high-resolution displays, making it perfect for complex visualization tasks. Additionally, it supports a variety of compute APIs, including CUDA, TensorRT, ONNX, DirectCompute, OpenCL, and OpenACC.Energy Efficiency and Thermal Management
Designed with high efficiency and performance in mind, the NVIDIA L40S consumes 350 W of power and features a passive cooling solution with bi-directional airflow support. This ensures optimal thermal management, even in the most demanding environments. The GPU operates within a temperature range of 0°C to 50°C and can be stored in temperatures ranging from -40°C to 75°C.Robust Software and API Support
The NVIDIA L40S supports a wide range of software and APIs, including NVIDIA RTX Virtual Workstation (vWS), NVIDIA Virtual PC (vPC), NVIDIA Virtual Applications (vApps), and NVIDIA AI Enterprise (NVAIE). It also supports graphics APIs such as DirectX 12.075, Shader Model 5.175, OpenGL 4.686, and Vulkan 1.186, making it a versatile choice for various applications.Reliability and Security
With ECC support and SR-IOV support, the NVIDIA L40S ensures data integrity and reliability. It is NEBS Level 3 ready, making it suitable for deployment in mission-critical environments. The GPU is designed to operate within a safe humidity range of 5% to 85% and can be stored in humidity conditions ranging from 5% to 95%.Ideal for AI, Performance Graphics, and VDI
The NVIDIA L40S Tensor Core GPU is specifically designed to handle AI, performance graphics, and VDI workloads. Its advanced features and robust performance make it an indispensable tool for IT professionals looking to enhance their computational capabilities.* Shown with sparsity. Specifications 1/2 lower without sparsity.
Compute Performance
Acceleration Applications: Machine Learning, Deep Learning, Virtual Desktops. Compute APIs: CUDA, TensorRT, ONNX, DirectCompute, OpenCL, OpenACC. Graphics APIs: DirectX 12.075, Shader Model 5.175, OpenGL 4.686, Vulkan 1.186.
| INT4 Performance | 733 TOPS | 1466 TOPS* |
|---|---|
| INT8 Performance | 733 TOPS | 1466 TOPS* |
| Mixed Precision | 183.2 TF | 366.4 TF* |
| Single Precision | 91.6 TF |
Features & Capabilities
vGPU Software: NVIDIA RTX Virtual Workstation (vWS), NVIDIA Virtual PC (vPC), NVIDIA Virtual Applications (vApps), NVIDIA AI Enterprise (NVAIE).
| Encoder/Decoder | 3 NVENC, 3 NVDEC (includes AV1 encode & decode) |
|---|---|
| Thermal Solution | Passive |

