PNY 2U SERVER SOLUTION
MOST COMPACT 2U SUPER NODE
The PNYSER28 server solution is a super-computing module designed and optimized for NVIDIA Tesla computing accelerators. This 2U rack offers the ability to receive up to 8 double slot GPU cards while the rest of the market offers 2U racks that can receive up to 6 cards at best in a 2U chassis.
With such power, this system is able to offer vast gains in processing time for the most supercomputing intensive applications, from cutting-edge rendering farms to the latest simulations in leading scientific research.

The 8 GPU cards are supported by a central backbone made of a dual socket platform supporting the Intel Xeon processors E5-2600 v3/v4 product family, dual GbE or dual 10G SFP+ LAN connections, and 8 x 2.5” hot-swappable HDD bays supporting SATA 6Gbs.
Indeed, the PNYSER28 series is a highly energy efficient system thanks to the innovative rack design and internal placement. Thanks to its power saving architecture, including high performance 80mm fans and a top efficiency 80+ Platinum level power supply, this product shows excellent airflow and thermal properties and manages to keep the power consumption at the lowest levels possible. This combination of large computing capability, high energy efficiency and compact 2U footprint allows the PNYSER28 series to deliver extreme megaflops per watt performance and therefore to help the most advanced supercomputing infrastructures to significantly drive their energy consumption down while boosting the execution of their applications.

The unprecedented computing density of PNYSER28 satisfies the needs of organizations seeking the largest amount of computing power and a minimal footprint at the same time.
RECOMMENDED CONFIGURATION |
CPU |
For HPC/Deep Learning: 2 x Xeon 8-Core E5-2620v4 2,1Ghz 20MB
For VDI: 2 x Xeon 12-Core E5-2687Wv4 3,0Ghz 30MB |
Memory |
For HPC/Deep Learning: 16 x 8GB DDR4 2400 LRDIMM ECC
For VDI: 16 x 32GB DDR4 2400 LRDIMM ECC |
Storage |
For HPC/Deep Learning: 1 x 2.5" 240GB + 1 x 2,5” 1.6TB SATA3 hot-swappable SSD
For VDI: 2 x 2.5" 240GB SATA3 hot-swappable SSD |
For HPC/Deep Learning |
GPU Cards |
4 x NVIDIA Tesla P100 |
4 x NVIDIA Tesla K80 |
4 x NVIDIA Tesla M40 |
Per GPU Card Specs |
4.7 Tflops Double Precision
GDDR5 memory: 16GB
NVIDIA CUDA Cores: 3584
732 GB/sec memory bandwidth |
2.91 Tflops Double Precision
GDDR5 memory: 24GB
NVIDIA CUDA Cores: 4992
480 GB/sec memory bandwidth |
7.0 Tflops Single Precision
GDDR5 memory: 24GB
NVIDIA CUDA Cores: 3072
288 GB/sec memory bandwidth |
For VDI |
GPU Cards |
2 x NVIDIA Tesla M10 |
2 x NVIDIA Tesla M60 |
Per GPU Card Specs |
4 NVIDIA Maxwell GPUs
Up to 64 concurrent users
GDDR5 memory: 32GB
NVIDIA CUDA Cores: 2560 |
2 NVIDIA Maxwell GPUs
Up to 32 concurrent users
GDDR5 memory: 16GB
NVIDIA CUDA Cores: 4096 |
HIGH PERFORMANCE COMPUTING (HPC)
Challenge in HPC:
Intensive and fast algorithmic computing, such as volume rendering, data-parallel vector calculation, multi-layer data linearity transformation, molecular fragments modeling.
GPU Computing as the Solution:

DEEP LEARNING
Challenge in Deep Learning:
High data training and validation iterations for linear and non-linear complex features with multiple processing layers and parameters; rapid object detection and parsing.
GPU Computing as the Solution:
VIRTUAL DESKTOP INFRASTRUCTURE (VDI)
Challenge in VDI:
Heavy graphics workload for the CPU; lengthy transfer delay of bulky graphics data; inefficient distribution of individual graphics resources; maximizing graphics performance at minimum latency.
GPU Computing as the Solution: