Wave’s Dataflow Server for Deep Learning

The Wave Computing dataflow server for datacenter environments is ideal for training and inferencing neural networks for demanding enterprise applications.

The Wave compute server is based on a revolutionary dataflow architecture that improves deep learning performance hundreds of times more than CPUs, GPUs and FPGAs, giving data scientists faster results and improved scaling in the datacenter. Each compute server comes in a 3U form factor that easily fits into existing datacenter environments.

With ONNX interoperability, Wave’s dataflow computer server can support a range of frameworks such as Tensorflow, Caffe, MXNet and more. Also, the Dataflow Processing Unit (DPU) based boards within each server are upgradeable, allowing for next-generation high-bandwidth memory clusters and future Wave DPUs.

 
WaveComputing_ProductMockup.png
 
WaveComputing_Datacenter_8.15.18_SpecTable.png

Wave Computing is revolutionizing AI and deep learning from the datacenter to the edge with its AI-native dataflow-based systems and embedded solutions. Enabling data scientists to experiment, develop, test, deploy and run AI applications at the speed of business, and more economically than ever before.

Interested in getting early access to Wave Computing’s pioneering AI system? Contact us today to see if you qualify!