NVIDIA Tesla V100S now Shipping in Exxact Servers

Whether your workload is training neural networks or creating powerful high-performance computing (HPC) simulations, Exxact’s TensorEX® line of servers featuring the NVIDIA V100S Tensor Core GPU offers the highest performance of any single GPU—enabling data scientists, researchers, and engineers to tackle challenges that were once unsolvable.

Exxact TensorEX servers can support up to 10x NVIDIA V100S Tensor Core GPUs in a single system, and are used by some of the top research institutions in the world, such as Carnegie Mellon, Stanford, and Stony Brook University.

Whether you are looking for a single GPU node or looking to scale out your infrastructure, Exxact TensorEX servers are ideal for meeting the ever-growing compute demands of GPU-accelerated applications.

The Tesla V100S is NVIDIA’s flagship data center GPU and when matched with Exxact’s TensorEX GPU server platforms, customers will be afforded great cost savings unmatched performance.

NVIDIA Tesla V100S - Exxact Corp

NVIDIA Tesla V100S Performance Specifications

Double-Precision Performance 8.2 teraFLOPS
Single-Precision Performance 16.4 teraFLOPS
Deep Learning Performance 130 teraFLOPS
Memory Capacity 32GB CoWoS Stacked HBM2
Memory Bandwidth 1134 GB/s
Max Power Consumption 250 Watts
Memory Bus Width 4096-bit
CUDA Cores 5120
Interface PCI Express 3.0 x16
Thermal Interface Passive Heatsink
Form Factor Dual Slot, Full Height
Display Connection None
Power Connection 8-pin Power Connector

The NVIDIA V100S GPU can operate in parallel on a single server or scale up to massive supercomputers and datacenters with hundreds or thousands of GPUs, making it the ideal choice for training deep learning models or running large HPC simulations in record time.

For more information on TensorEX GPU systems featuring NVIDIA GPUs, visit: https://www.exxactcorp.com/NVIDIA-GPU-HPC-Clusters or contact sales@exxactcorp.com


Back to Blog Homepage