DEEP LEARNING PLATFORMS POWERED BY PNY

The Most Powerful GPU Servers & Workstations Optimized for DL Workloads

DL-E48X

Configurable PCIe Root Complex Deep Learning Server

The first-in-industry accelerated 8x GPU computing solution to feature re-configurable single & dual root complex PCIe architecture, allowing for hardware optimization on the fly for AI and DL training, inference, HPC compute, rendering and virtualization applications.

Available PNY cards: GV100, GP100, P5000 and P6000.

ServMax™ HGX-1

Hyperscale GPU Accelerated AI Cloud Solution

Powered by NVIDIA Tesla GPUs and NVLink™ high-speed interconnect technology, the HGX-1 is purpose-built for AI/HPC cloud computing. Hosting 8x Tesla V100 SXM2 GPUs in a 4U chassis, the HGX-1 features 5,120 Tensor Cores and 40,960 CUDA Cores for 1 PFLOP of Tensor Core operations, and 125 TFLOPS of SP performance.

Available PNY cards: GV100, GP100, P5000 and P6000.

HGX-1

GPU Accelerated AI Cloud Solution

Powered by NVIDIA Tesla GPUs and NVLink™ high-speed interconnect technology, the HGX-1 is purpose-built for AI/HPC cloud computing. Hosting 8x Tesla V100 SXM2 GPUs in a 4U chassis, the HGX-1 features 5,120 Tensor Cores and 40,960 CUDA Cores for 1 PFLOP of Tensor Core operations, and 125 TFLOPS of SP performance.

Available PNY cards: GV100, GP100, P5000 and P6000.

DL-E280

Ultra-Compact High-Density Deep Learning Server

8x GPU Machine Learning server with Intel Xeon Scalable Processor delivering up to 112 TFLOPs of single precision when populated with NVIDIA Tesla V100 GPU accelerators.

Available PNY cards: GV100, GP100, P5000 and P6000.

ServMax™ P47

AMD EPYC GPU-Accelerated Compute Platform

Featuring the latest AMD EPYC™ CPUs, ServMax™ P47 supports both NVIDIA Tesla GPUs and Vega-based Radeon Instinct™ GPUs. The single socket platform easily manages machine learning, advanced rendering and HPC workloads previously only possible on high-end dual socket systems, making it a superior alternative to existing GPU solutions in the market.

ServMax™ P47

AMD EPYC GPU-Accelerated Compute Platform

Featuring the latest AMD EPYC™ CPUs, ServMax™ P47 supports both NVIDIA Tesla GPUs and Vega-based Radeon Instinct™ GPUs. The single socket platform easily manages machine learning, advanced rendering and HPC workloads previously only possible on high-end dual socket systems, making it a superior alternative to existing GPU solutions in the market.

DL-E140

Ultra-Compact 1U 4x GPU Deep Learning Server

High-density, ultra-compact building block for GPU-powered Deep Learning and HPC clusters, accelerated by Intel Xeon Scalable Processors with NVIDIA® GPUs, including Tesla V100, P100, P40, Quadro P5000 and P6000.

Available PNY cards: GV100, GP100, P5000 and P6000.

DL-E200

Mini Deep Learning Workstation

Ultra-compact high-end Deep Learning development workstation, perfect for AI startups and labs. This ultra-quiet micro ATX workstation features 2x NVIDIA GTX Titan V or 1080Ti GPUs, on-board dual 10G Ethernet and server grade motherboard.

Available PNY cards: GV100, GP100, P5000, P6000 and 1080Ti.

DL-E400

High-Performance Deep Learning DevBox

Our best-selling Deep Learning workstation for Deep Learning development! This ultra-quiet compact workstation featuring 4x NVIDIA GTX Titan V or 1080Ti GPUs, on-board dual 1G/10G Ethernet and workstation-grade motherboard.

Available PNY cards: GV100, GP100, P5000, P6000 and 1080Ti.

[SMART]Rack AI

Turnkey High Performance Machine Learning Cluster

Rackscale solution featuring up to 96x NVIDIA® Tesla GPU accelerator cards for up to 1.34 PFLOPs per rack. All-Flash storage for an ultra-fast in-rack data repository, dual 25GbE ROCE or 100G EDR InfiniBand high speed networking, [SMART]DC Data Center Manager and optional in-rack battery for power loss protection.

Available PNY cards: GV100, GP100, P5000 and P6000.

SPECIFICATIONS

GTX 1080GTX 1080TiNVIDIA Titan XNVIDIA Titan XPPNY Quadro P6000NVIDIA Tesla P40

Micro Architecture

Pascal

Pascal

Pascal

Pascal

Pascal

Pascal

Chipset

Pascal GP104-400

Pascal GP104102

Pascal GP104102

Pascal GP104102

Pascal GP104102

Pascal GP104102

Application

Consumer

Consumer

Enthusiast

Enthusiast

Workstation/Server

Server

Thermal Solution

Active

Active

Active

Active

Active

Passive

Cuda Cores

2560

3854

3854

3840

3840

3840

Peak SP Performance

9.0 TFLOPS

11.4 TFLOPS

12.0 TFLOPS

12.0 TFLOPS

12.0 TFLOPS

12.0 TFLOPS

GPU Memory

8GB GDDR5X

11GB GDDR5X

12GB GDDR5X

12GB GDDR5X

24GB GDDR5X

24GB GDDR5X

Memory Interface

256-bit

352-bit

384-bit

384-bit

384-bit

384-bit

Memory Bandwidth

320GB/s

484GB/s

480GB/s

548GB/s

432GB/s

346GB/s

System Interface

PCIe 3.0 x 16

PCIe 3.0 x 16

PCIe 3.0 x 16

PCIe 3.0 x 16

PCIe 3.0 x 16

PCIe 3.0 x 16

Max Power Consumption

180W

250W

250W

250W

250W

250W

Live Chat