NVIDIA® Tesla™ M40 GPU Computing Accelerator - 12GB GDDR5 - Passive Cooler

NVIDIA part #: 900-2G600-0000-000

Tesla M40 GPU Accelerator
Deep learning is the fundamental building block for modern AI. The NVIDIA Tesla M40 is the world’s fastest accelerator for deep learning training in the datacenter, purpose-built to deliver models within hours versus days on CPU-only servers. Featuring Maxwell, the most advanced GPU architecture, and large-capacity memory design, Tesla M40 enables customers to iterate quickly and deploy increasingly complex models faster. It is the newest addition to the NVIDIA Tesla accelerated Computing Platform, which enables deep learning practitioners leverage cuDNN, DIGITS, and popular frameworks optimized for M40 to easily build, test, and deploy modern AI applications in the datacenter.

Price: $4,499.00

Add To Order
Main Specifications
Product SeriesTesla M40
Host InterfacePCI Express 3.0 x16
GPU ArchitectureMaxwell
Detailed Specifications
Streaming Processor Cores3072 CUDA Cores
Memory Interface384-bit
Max Memory Size12 GB GDDR5
Max Memory Bandwidth288 GB/s
Peak Double Precision floating point performance (GFLOP)0.2 Teraflops
Peak Single Precision floating point performance (GFLOP)7 Teraflops with NVIDIA GPU Boost
NVIDIA CUDA™ TechnologyYes
Dual SlotDual Slot, Full Height
Dimensions4.4” H × 10.5” L
Supplementary Power Connectors1x 8 Pin
Max Graphics Card Power (W)250W