TESLA A40 RTX PASSIVE 300W 48GB

Description

NVIDIA Ampere Architecture with CUDA Core

Dual-speed processing for single-precision floating-point operations (FP32) and improved power efficiency offer significant performance improvements for graphics and simulation workflows such as 3D CAD (computer-aided design) and CAE (computer-aided engineering).

Second-generation RT Core

With 2X the speed of the previous generation and the ability to simultaneously perform ray tracing with shading or denoising capabilities, the second-generation RT Cores provide massive accelerations for workloads such as photorealistic rendering of film content, architectural design evaluations, and virtual prototyping of product designs. This technology also accelerates motion blur rendering with ray tracing for faster results with greater visual accuracy.

Third-generation Tensor Core

The new precision Tensor Float 32 (TF32) provides 5 times faster training speed than the previous generation to accelerate the training of AI and data science models without the need for code changes. Hardware support for structural density doubles the throughput for inference. Tensor Cores also bring AI to graphics with features such as DLSS, AI-based denoising, and advanced editing for some applications.

48GB of GPU memory

Ultra-fast GDDR6 memory, scalable up to 96GB with NVLink, gives data scientists, engineers and creative professionals all the space they need to work with huge datasets and workloads such as data science and simulation.

Third-generation NVIDIA NVLink

Connects up to two A40 GPUs together to scale from 48GB of GPU memory to 96GB. Increased GPU-GPU interconnect bandwidth provides a single scalable memory to accelerate graphics and compute workloads and address more voluminous data sets. A new, more compact NVLink connector enables functionality across a wider range of servers.

Ready for virtualization

Next-generation enhancements with NVIDIA Virtual GPU (vGPU) software enable the creation of larger, more powerful virtual workstation instances for remote users, enabling high-end remote design, AI, and compute workloads.

PCI Express Gen 4

PCI Express Gen 4 doubles the bandwidth over PCIe Gen 3, improving data transfer rates from CPU memory for intensive tasks such as AI, data science, and 3D design. Faster PCIe performance also accelerates GPU direct memory access (DMA) transfers, enabling faster I/O communication of video data between the GPU and GPUDirect for Video compatible devices, providing a powerful solution for live broadcasting. The A40 is backward compatible with PCI Express Gen 3 ensuring flexibility in deployment.

Data center efficiency and security

With a dual-slot, energy-efficient design, the NVIDIA A40 is 2x more efficient than the previous generation and is designed to fit a wide range of workstations from OEMs around the world. NVIDIA A40 also provides secure and measured boot with reliable hardware root, which ensures that the firmware has not been tampered with or corrupted.