NVIDIA L40 Tensor Core GPU 48GB GDDR6 AI VDI Graphics 900-2G150-0000-000
Experience unparalleled AI and VDI performance with the NVIDIA L40 Tensor Core GPU. Featuring the advanced NVIDIA Ada Lovelace architecture, this powerhouse boasts 48GB of GDDR6 memory, 568 Tensor Cores, and 18,176 CUDA Cores. With a base clock speed of 735 MHz and a maximum boost clock speed of 2490 MHz, it delivers exceptional single-precision performance of 90.5 TF. Designed for high efficiency and performance, it supports PCIe 4.0 x16 interface, ECC, SR-IOV, and a passive cooling solution. Ideal for machine learning, deep learning, and virtual desktops, it ensures seamless integration with NVIDIA's vGPU software suite.
What to Expect from Enterasource
Why choose Enterasource?
At Enterasource, we pride ourselves on delivering top-notch service to our valued customers. Here are just a few of the benefits you can expect when choosing us for your IT needs
Complimentary ground shipping on orders over $50 throughout the contiguous United States (exclusions apply to Hawaii and Alaska).
Efficient same-day shipping for in-stock orders placed before 3PM PST.
Rapid one business day quote response time.
Industry-leading next business day parts
replacement warranties to ensure your equipment stays up and running.
Our commitment to exceptional customer service and timely delivery make us the ideal choice for all your IT hardware needs. Choose Enterasource for a hassle-free experience and the highest quality product on the market.
Part Type | GPU |
---|---|
Description | NVIDIA L40 Tensor Core GPU 48GB GDDR6 AI VDI Graphics 900-2G150-0000-000: Unleash Unprecedented PerformanceEnterasource proudly presents the NVIDIA L40 Tensor Core GPU, a powerhouse designed to meet the rigorous demands of modern IT environments. With its cutting-edge NVIDIA Ada Lovelace architecture and a massive 48GB of GDDR6 memory, this GPU is engineered to deliver unparalleled performance for AI, machine learning, and virtual desktop infrastructure (VDI) applications. Whether you're an IT professional managing large-scale data centers or a researcher pushing the boundaries of AI, the NVIDIA L40 is your ultimate solution. Key Features and Benefits of the NVIDIA L40 Tensor Core GPU 48GB GDDR6 AI VDI Graphics 900-2G150-0000-000Revolutionary PerformanceEquipped with 18176 CUDA cores and 568 Tensor cores, the NVIDIA L40 offers a staggering single-precision performance of 90.5 TF. For mixed-precision tasks, it delivers up to 362.1 TF*, making it ideal for complex computations and deep learning models. The GPU's base clock speed of 735 MHz and a maximum boost clock speed of 2490 MHz ensure that you get the best performance under varying workloads. Advanced Memory and BandwidthThe 48GB GDDR6 memory provides ample space for large datasets, while the PCIe 4.0 x16 interface ensures a high interconnect bandwidth of 864 GB/s. This combination allows for faster data transfer and reduced latency, crucial for real-time applications and high-performance computing tasks. Energy Efficiency and Thermal ManagementDesigned with high energy efficiency, the NVIDIA L40 consumes only 300 W of power, making it a cost-effective choice for large-scale deployments. The passive cooling solution and bi-directional airflow support ensure optimal thermal management, maintaining an operating temperature range of 0°C to 50°C. Robust Compatibility and SupportThe NVIDIA L40 is compatible with various server platforms supporting NVIDIA GPUs, making it a versatile addition to your IT infrastructure. It supports a wide range of compute and graphics APIs, including CUDA, TensorRT, DirectX 12.075, and Vulkan 1.186. Additionally, it offers robust vGPU software support, including NVIDIA RTX Virtual Workstation (vWS) and NVIDIA AI Enterprise (NVAIE). Enhanced Security and ReliabilityWith ECC support and SR-IOV support, the NVIDIA L40 ensures data integrity and secure multi-tenant environments. Its NEBS Level 3 readiness further guarantees reliability in demanding conditions, making it a trusted choice for mission-critical applications. Versatile Application SupportWhether you're working on machine learning, deep learning, or virtual desktops, the NVIDIA L40 is designed to accelerate your workloads. Its encoder/decoder support, including AV1 encode and decode, further enhances its versatility, making it suitable for a wide range of applications. * Shown with sparsity. Specifications 1/2 lower without sparsity. |
Manufacturer | NVIDIA |
MPN | 900-2G150-0000-000 |
Alternate Part Numbers | 699-2G150-0000-xxx |
Condition | Refurbished |
Display Connectors | 0 |
Display Ports | 4 |
HDMI Ports | 0 |
GPU Architecture | NVIDIA Ada Lovelace |
Memory Type | GDDR6 |
Base Clock Speed | 735 MHz |
Maximum Boost Clock Speed | 2490 MHz |
UEFI Support | Yes |
Tensor Cores | 568 |
CUDA Cores | 18176 |
RT Cores | 142 |
Single-Precision Performance | 90.5 TF |
Mixed-Precision Performance | 181.05 TF | 362.1 TF* |
INT8 Performance | 362 TOPS | 724 TOPS* |
INT4 Performance | 724 TOPS | 1448 TOPS* |
Interface | PCIe 4.0 x16 |
Interconnect Bandwidth | 864 GB/s |
Power Consumption | 300 W |
Max TDP Power | 300 W |
Power Connector | 16-pin |
Form Factor | Dual Slot, Full Height |
SR-IOV Support | Yes |
ECC Support | Yes |
Energy Efficiency | High efficiency and performance |
Thermal Solution | Passive |
Cooling Type | Passive |
Airflow Support | Bi-directional |
Operating Temperature | 0 °C to 50 °C |
Storage Temperature | -40 °C to 75 °C |
Safe Humidity Range While Operating | 5% to 85% |
Humidity Storage Range | 5% to 95% |
Compute APIs | CUDA, TensorRT, ONNX, DirectCompute, OpenCL, OpenACC |
Acceleration Applications | Machine Learning, Deep Learning, Virtual Desktops |
Encoder/Decoder Support | 3 NVENC, 3 NVDEC (includes AV1 encode & decode) |
vGPU Software Support | NVIDIA RTX Virtual Workstation (vWS), NVIDIA Virtual PC (vPC), NVIDIA Virtual Applications (vApps), NVIDIA AI Enterprise (NVAIE) |
Graphics APIs | DirectX 12.075, Shader Model 5.175, OpenGL 4.686, Vulkan 1.186 |
NEBS Ready | Level 3 |
MIG Support | No |
Workload | AI/Performance graphics/VDI |
Compatibility | Compatible with various server platforms supporting NVIDIA GPUs |