Call us9940555925|WhatsApp7395808630
Redington
NVIDIA A40 48GB PCIe Non-CEC Accelerator for HPE
NVIDIA A40 48GB PCIe Non-CEC Accelerator for HPE
NVIDIA A40 48GB PCIe Non-CEC Accelerator for HPE
HPE

NVIDIA A40 48GB PCIe Non-CEC Accelerator for HPE

Components
MPN:R9S37C
SKU:CMPLDX0018

Product Highlights

  • NVIDIA A40 48GB PCIe Non-CEC Accelerator for HPE

Specifications

General

Brand

HPE

Compatible With

Compatible with HPE SY480 Gen10 E2,HPE Apollo 6500 XL645d Gen10 Plus,HPE Apollo 6500 XL675d Gen10 Plus,HPE DL380 Gen10 Plus,HPE DL385 Gen10 Plus v2,HPE SY480 Gen10 Plus E2,HPE DL380 Gen10,HPE DL580 Gen10,HPE Superdome Flex 280,HPE Edgeline EL8000 e920d/e920t

Country Of Origin

Singapore

Form Factor

Plug-in Card

Product Type

Components

Processor

Chipset

NVIDIA A40

Memory

Memory

8 GB

Ports

Platform Supported

HPE ProLiant Rack/Tower/BladeSystem Servers/Synergy/HPE Specialized Compute Platforms

Physical

Dimension

3.47 x 26.7 x 11.1 cm

Weight-Kg

0.99 kg

Product Manuals

No manuals available for this product

Description

Product Details

  • - Graphics processing unit (GPU) accelerator based on NVIDIA Turing architecture
  • - Provides 4,096 CUDA cores and 48GB of 8008-MHz GDDR6 memory
  • - Delivers up to 26 Teraflops of mixed precision and 8.2 Teraflops of double precision performance
  • - Features NVLink and PCIe Gen3 interface for fast data transfer
  • -Certified to work in NVIDIA-ONTM enabled HPE ProLiant, Synergy and HPC servers
  • - Supports compute-intensive workloads like Deep Learning, Data Science, HPC and Graphics
  • - Provides NVIDIA NVSwitch for scaling multi-GPU solutions
  • - Features Tensor Cores for deep learning training and inference acceleration
  • In summary, the NVIDIA A40 GPU accelerator for HPE delivers top-tier deeplearning and HPC performance to accelerate AI, data science andengineering applications running on HPE servers. With its 48GB ofhigh-bandwidth memory, massive CUDA core count and Tensor Cores,it is designed to boost the performance of the most demanding workloads.