NVIDIA A40

NVIDIA A40

Unleash next-gen visualisation in your data center with the NVIDIA A40 GPU

NVIDIA A40

Unlock the power of next-generation visualisation with the NVIDIA A40 data center GPU. Experience unparalleled performance for demanding professional workloads, virtual workstations, and more. NVIDIA's A40 GPU boosts your creativity and productivity with the latest Ampere architecture, massive memory capacity, and cutting-edge artificial intelligence (AI) features.

Key Features

  • Unleash Ampere Power - Experience double the performance for single-precision floating-point operations and improved power efficiency compared to the previous generation, thanks to the A40's cutting-edge NVIDIA Ampere architecture.
  • Accelerate workloads - Handle massive datasets and complex workloads with ease, thanks to 48GB of GDDR6 memory and dedicated processing cores for ray tracing, AI, and visualisation.
  • Scale with confidence - Deploy the NVIDIA A40 in virtualised environments and leverage NVLink for seamless multi-GPU scaling, ensuring your data center stays flexible and adaptable to evolving needs.

Specifications

Architecture Ampere
GPU memory 48GB GDDR6 with ECC
Memory Bandwidth 696GB/s
Interconnect Interface NVIDIA® NVLINK® 112.5GB/s (bi-directional)
Connecting two NVIDIA A40 cards with NVLink® to scale performance and memory capacity to 96 GB is only possible if your application supports NVLink® technology
PCIe Gen4: 64GB/s
NVIDIA Ampere Architecture-based CUDA Cores 10,752
NVIDIA 2nd Gen RT Cores 84
NVIDIA 3rd Gen Tensor Cores 336
Peak FP32 TFLOPS (non-Tensor) 37.4
Peak FP16 Tensor TFLOPS with FP16 Accumulate 149.7 || 299.4 with sparsity
Peak TF32 Tensor TFLOPS 74.8 || 149.6 with sparsity
RT Core Performance TFLOPS 73.1
Peak BG16 Tensor TFLOPS with FP32 Accumulate 149.7 || 299.4 with sparsity
Peak INT8 Tensor TOPS 299.3 || 598.6 with sparsity
Peak INT4 Tensor TOPS 598.7 || 1,197.4 with sparsity
Form Factor Dual-slot || 4.4" (H) x 10.5" (L)
Display Ports 3 * DisplayPort 1.4
Supports NVIDIA Mosaic and Quadro® Sync
The A40 is configured for virtualisation by default with physical display connectors disabled. Enabling this function can be done through the management software tools.
Max Power Consumption 300W
Power connector 8-pin CPU
Thermal Solution Passive
Virtual GPU (vGPU) Software Support NVIDIA vPC/vApps
NVIDIA RTX Virtual Workstation
NVIDIA Virtual Compute Server
vGPU Profiles Supported View supported profiles at: https://www.nvidia.com/content/dam/en-zz/Solutions/design-visualization/solutions/resources/documents1/Virtual-GPU-Packaging-and-Licensing-Guide.pdf
Decode/Encode 1 * NVENC
2 * NVDEC
Includes AV1 decode
Secure and Measured Boot with Hardware Root of Trust Yes (optional)
NEBS Ready Level 3
Compute APIs CUDA
DirectCompute
OpenCL™
OpenACC®
Graphics APIs DirectX 12.07 - Supports DX 12.0 API, Hardware Feature Level 12 + 1
Shader Model 5.17 - Supports DX 12.0 API, Hardware Feature Level 12 + 1
OpenGL 4.68 - Based on Khronos specifications. Learn more at: www.krhonos.org/conformance.
Vulkan 1.18 - Based on Khronos specifications.
MIG Support No