NVIDIA L40 GPU
Delivering unprecedented visual computing performance for the data center.
Introduction
Enterprises require data center technology that delivers extreme performance and scale for complex visual computing and scientific workflows. The NVIDIA L40 GPU offers next-generation graphics, compute, and AI capabilities, built on the NVIDIA Ada Lovelace architecture. It features the latest generation RT, Tensor, and CUDA cores for groundbreaking visualization and compute performance.
Powered by the NVIDIA Ada Lovelace Architecture
Third-Generation RT Cores
Provide enhanced throughput and concurrent ray-tracing and shading for improved ray-tracing performance. Accelerates renders for product design, architecture, engineering, and construction workflows. Enables lifelike designs with hardware-accelerated motion blur for real-time animations.
Fourth-Generation Tensor Cores
Offer hardware support for structural sparsity and optimized TF32 format for faster AI and data science model training. Accelerates AI-enhanced graphics, including DLSS, for upscaled resolution and better performance.
Large GPU Memory
Features 48GB of ultra-fast GDDR6 memory for memory-intensive applications like data science, simulation, 3D modeling, and rendering. Supports vGPU software for allocating memory to multiple users across creative, data science, and design teams.
Data-Center Ready
Designed for 24x7 enterprise data center operations with power-efficient hardware. Optimized for scale and maximum performance. Includes secure boot with root of trust technology for enhanced security and is NEBS Level 3 compliant. Features a dual-slot, passively cooled design and is available in NVIDIA-Certified Systems™ from leading OEMs.
Accelerate Next-Generation Workloads
- NVIDIA Omniverse™ Enterprise
- Rendering and 3D Graphics
- Generative AI and Inference
- High-Performance Virtual Workstations with NVIDIA RTX™ Virtual Workstation (RTX vWS) Software
- AI Training and Data Science
- Streaming and Video Content
Technical Specifications
| Specification | Details |
|---|---|
| GPU Architecture | NVIDIA Ada Lovelace architecture |
| GPU Memory | 48GB GDDR6 with ECC |
| Memory Bandwidth | 864GB/s |
| Interconnect Interface | PCIe Gen4x16: 64GB/s bi-directional |
| NVIDIA Ada Lovelace architecture-based CUDA Cores | 18,176 |
| NVIDIA third-generation RT Cores | 142 |
| NVIDIA fourth-generation Tensor Cores | 568 |
| RT Core performance TFLOPS | 209 |
| FP32 TFLOPS | 90.5 |
| TF32 Tensor Core TFLOPS | 90.5 | 181** |
| BFLOAT16 Tensor Core TFLOPS | 181.05 | 362.1** |
| FP16 Tensor Core | 181.05 | 362.1** |
| FP8 Tensor Core | 362 | 724** |
| Peak INT8 Tensor TOPS | 362 | 724** |
| Peak INT4 Tensor TOPS | 724 | 1448** |
| Form Factor | 4.4" (H) x 10.5" (L) - dual slot |
| Display Ports | 4 x DisplayPort 1.4a |
| Max Power Consumption | 300W |
| Power Connector | 16-pin |
| Thermal | Passive |
| Virtual GPU (vGPU) software support | Yes |
| VGPU Profiles Supported | See Virtual GPU Licensing Guide |
| NVENC I NVDEC | 3x | 3x (Includes AV1 Encode and Decode) |
| Secure Boot with Root of Trust | Yes |
| NEBS Ready | Level 3 |
| MIG Support | No |
| NVLink Support | No |
** With Sparsity.
Ready to get started?
To learn more about the NVIDIA L40 GPU, visit: www.nvidia.com/L40
