Description

Do you need a cutting-edge graphics card solution for visual computing workflows in your data center? NVIDIA A40 48GB GDDR6 is a powerful and versatile data center solution that also excels in efficiently catering to the next-gen challenges faced by artists, designers, engineers, and scientists. Experience double-speed processing with A40 GPU’s NVIDIA Ampere architecture, combining 2nd Gen RT Cores and 3rd Gen Tensor Cores. While the RT cores support complex CAD, CAE, ray tracing, and photorealistic rendering tasks, the TF32 precision provides improved training throughput to speed up data science and AI models. The massive 48GB GDDR6 VRAM and a PCIe 4.0 interface account for much-improved transmission speeds in data-intensive tasks. A40 is a compact dual-slot graphics card with a power-efficient design compatible with most known and popular servers. Our services are available across Oman with free shipping and a 1-year warranty. Upgrade your data center with NVIDIA A40 featuring the CEC 1712 chipset for enhanced security. Call us to know more.

Technical Specifications

Brand	NVIDIA
GPU architecture	NVIDIA Ampere architecture
Gigabytes of memory	48GB
Memory type	GDDR6
CUDA Cores	10752
Memory bandwidth	696 GB/s
Tensor Cores	336
RT Cores	84
Max Power Consumption	300 W
Graphics Bus	PCI Express Gen 4 x 16
Peak FP32 Tensor TFLOPS	74.8/149.6
Peak FP16 Tensor TFLOPS	149.7/299.4
Peak INT8 Tensor TOPS	299.8/599.6
Peak INT4 Tensor TOPS	599.7/1199.4
Thermal solution	Passive
NEBS Ready	Level 3
Power Connector	8-pin CPU
NVLink	2-way low profile (2-slot)
vGPU Software Support	NVIDIA vPC/vApps, NVIDIA RTX Virtual Workstation, NVIDIA Virtual Compute Server
Display Connectors	3x DisplayPort 1.4, Off by Default
Compute APIs	CUDA, DirectCompute, OpenCL, OpenACC
Graphics APIs	DirectX 12.075, Shader Model 5.175,OpenGL 4.686, Vulkan 1.18